What HW/SW issues influence org/Emacs performance?

What HW/SW issues influence org/Emacs performance?

[Karl Voit]

Hi!

I was thinking about which hardware and software attributes are the
most important factors for improving performance of Emacs and/or
Org-mode. The things I want to improve are 

(1) startup time of GNU Emacs (nice to have, no priority)

(2) using Org-mode agenda with approx. 20 Org-files 

(3) using agenda view including archive files 

The Org-mode files I regularly use in (2) contain 15,000 to 25,000
lines. The org_archive files contain more lines, three of them
40,000 to 130,000 lines with PROPERTY-drawers[1].

The biggest performance increase I'd like to see in moving in my
normal agenda (approx. 2-3s from one day to the next) and in my
agenda including the archive files (approx. 4-8s from one day to the
next). Usually, it is worth waiting :-)

On my system [3] I notice 100% load on one of my two cores when
Emacs is generating the agenda views. So I am not sure whether more
cores could increase the performance. Probably a faster CPU? Is
there something that makes Emacs use multiple cores in parallel?

We did some performance testing [1] for Memacs [2] but this only
tested a static situation on given platforms.

What do you think are the most important factors?

- CPU speed (of one core)
- type of CPU
- # CPU cores
- 32bit kernel vs. 64bit kernel
- faster I/O (through HDD->SSD or even faster SSD)
- choice of the file system (optimized for something? what?)
- operating system 
- possible ELISP optimizations (are there any besides compiling?)

Probably you already have an idea of a system which is designed to
be optimized for Emacs/Org-mode/Agenda ...? How does it look like?
:-)


... very curious about your thoughts ...

  1. https://github.com/novoid/Memacs/blob/master/docs/performance.org
  2. https://github.com/novoid/Memacs
  3. lenovo X200s, Core 2 Duo, SSD, Ubuntu 11.04, Gnome 2, GNU Emacs 23
-- 
Karl Voit

Re: What HW/SW issues influence org/Emacs performance?

[Achim Gratz]

Karl Voit writes:
> On my system [3] I notice 100% load on one of my two cores when
> Emacs is generating the agenda views. So I am not sure whether more
> cores could increase the performance. Probably a faster CPU? Is
> there something that makes Emacs use multiple cores in parallel?

No, Emacs is strictly single-threaded and will probably remain that way
for quite some time.  Note that there is talk to add concurrency support
to Emacs, but even then only one of those "Emacs threads" can run at any
given instant.  So this will help if some of those threads would
otherwise busy-wait on something, but not if there is a compute-bound
thread.  Nobody wants to open the can of worms that a truly
multithreaded ELisp is…


Regards,
Achim.
-- 
+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

Samples for the Waldorf Blofeld:
http://Synth.Stromeko.net/Downloads.html#BlofeldSamplesExtra

Re: What HW/SW issues influence org/Emacs performance?

[Luis Anaya]

Karl Voit <devnull@Karl-Voit.at> writes:

> I was thinking about which hardware and software attributes are the
> most important factors for improving performance of Emacs and/or
> Org-mode. The things I want to improve are 

It is an interesting question, but I think that it is not restricted to
Emacs or Org. Without going through a software profiling exercise, it is
just trying to shoot in the dark. In your case, you'll have a set of
relatively large files that will take some time to process. This will
cause the CPU to be bound during that time. Being that emacs is single
threaded, it's only one core that gets bound. 


>
> What do you think are the most important factors?
>
> - CPU speed (of one core)
> - type of CPU

I run Emacs on different platforms, but most of my development is done
on ARM (A pogo plug running Arch).  Things happen a bit slower, but it
is not unusable. The biggest factor on these systems is that their I/O 
performance is not great, but it only matters when the system is
executing a lot of processes that that affects the overall system load. 
I would not say that it is the type, but the speed and the factors that 
contribute your system load.


> - 32bit kernel vs. 64bit kernel

This only affects amount of available memory. Unless you are loading
files that are larger than 2 Gigabytes, it does not matter.

> - faster I/O (through HDD->SSD or even faster SSD)

Only if the system is in high I/O utilization. When the system is
relatively idle, it really does not matter being that it's dedicated to
the CPU. So... in this specific case in Emacs, it is the CPU speed that has the
largest impact on the overall performance. 


My 2 cents. 

-- 
Luis Anaya
papo anaya aroba hot mail punto com
"Do not use 100 words if you can say it in 10" - Yamamoto Tsunetomo

Wish: switch active time-stamps to inactive when CANCELED

[Karl Voit]

Hi!

I tend to keep events that I do not attend in my calendar for
archiving reasons. But I switch their status to CANCELED.
Unfortunately, those events still appear on my Google calendar and
so forth. I get my usual reminders and it messes up my free/busy
list.

Therefore, I am thinking that canceled events should turn their
active time-stamps into inactive with an Org-mode option. This way,
I can recall them in my agenda by typing ']' (show inactive
time-stamps) but they do not interfere with my daily workflow.

What do you think of such an option for Org-mode?

-- 
Karl Voit

Re: Wish: switch active time-stamps to inactive when CANCELED

[Daimrod]

[-- Attachment #1: Type: text/plain, Size: 1082 bytes --]

Karl Voit <devnull@Karl-Voit.at> writes:

> Hi!

Hi,

> I tend to keep events that I do not attend in my calendar for
> archiving reasons. But I switch their status to CANCELED.
> Unfortunately, those events still appear on my Google calendar and
> so forth. I get my usual reminders and it messes up my free/busy
> list.
>
> Therefore, I am thinking that canceled events should turn their
> active time-stamps into inactive with an Org-mode option. This way,
> I can recall them in my agenda by typing ']' (show inactive
> time-stamps) but they do not interfere with my daily workflow.
>
> What do you think of such an option for Org-mode?

Org Mode has already facilities for archiving.

You can either tag a headline and its children with the `archive' tag
(bound to C-c C-x a) or archive a headline and its children in a
separated file (bound to C-c $).

In both cases they won't be displayed in the Agenda unless you active
them (v a or v A).

See (info "(org) Archiving") for a complete description of Archiving in
Org Mode.

-- 
Daimrod/Greg

[-- Attachment #2: Type: application/pgp-signature, Size: 835 bytes --]

Re: Wish: switch active time-stamps to inactive when CANCELED

[Karl Voit]

* Daimrod <daimrod@gmail.com> wrote:
>
> Org Mode has already facilities for archiving.

Thanks for the pointer but I know and I am using archiving in
Org-mode.

> In both cases they won't be displayed in the Agenda unless you active
> them (v a or v A).

Right.

But in my workflow, I only archive whole projects when they are
finished completely. For multiple reasons, I don't want to separate
single events from the projects upfront.

I should have mentioned that in my first post, yes.

-- 
Karl Voit

Re: Wish: switch active time-stamps to inactive when CANCELED

[Daimrod]

[-- Attachment #1: Type: text/plain, Size: 746 bytes --]

Karl Voit <devnull@Karl-Voit.at> writes:

> * Daimrod <daimrod@gmail.com> wrote:
>>
>> Org Mode has already facilities for archiving.
>
> Thanks for the pointer but I know and I am using archiving in
> Org-mode.
>
>> In both cases they won't be displayed in the Agenda unless you active
>> them (v a or v A).
>
> Right.
>
> But in my workflow, I only archive whole projects when they are
> finished completely. For multiple reasons, I don't want to separate
> single events from the projects upfront.

Why doesn't the method with :ARCHIVE: tag fit your needs?

You can mark a subtree to ignore it. This way you keep the whole project
as a whole.

> I should have mentioned that in my first post, yes.

-- 
Daimrod/Greg

[-- Attachment #2: Type: application/pgp-signature, Size: 835 bytes --]

Re: Wish: switch active time-stamps to inactive when CANCELED

[Karl Voit]

* Daimrod <daimrod@gmail.com> wrote:
>
> Why doesn't the method with :ARCHIVE: tag fit your needs?
>
> You can mark a subtree to ignore it. This way you keep the whole project
> as a whole.

Good point indeed.

But still: there are two steps I have to make: (a) mark as canceled
because I want this "CANCELED" string attached and (b) add the
ARCHIVE tag.

Without "CANCELED", I would not interpret the event as not attended
by me.

-- 
Karl Voit

Re: Wish: switch active time-stamps to inactive when CANCELED

[Viktor Rosenfeld]

Hi,

Karl Voit wrote:

> But still: there are two steps I have to make: (a) mark as canceled
> because I want this "CANCELED" string attached and (b) add the
> ARCHIVE tag.

You can set the ARCHIVE tag when switching the TODO state automatically
using org-todo-state-tags-triggers.

For example (untestet):

#+BEGIN_SRC emacs-lisp
(setq org-todo-state-tags-triggers '(("CANCELLED" ("ARCHIVE" . t))))
#+END_SRC

Cheers,
Viktor

Re: Wish: switch active time-stamps to inactive when CANCELED

[Daimrod]

[-- Attachment #1: Type: text/plain, Size: 598 bytes --]

Viktor Rosenfeld <listuser36@gmail.com> writes:

> Hi,
>
> Karl Voit wrote:
>
>> But still: there are two steps I have to make: (a) mark as canceled
>> because I want this "CANCELED" string attached and (b) add the
>> ARCHIVE tag.
>
> You can set the ARCHIVE tag when switching the TODO state automatically
> using org-todo-state-tags-triggers.
>
> For example (untestet):
>
> #+BEGIN_SRC emacs-lisp
> (setq org-todo-state-tags-triggers '(("CANCELLED" ("ARCHIVE" . t))))
> #+END_SRC

Woo, I didn't know this one, that's a nice trick!

> Cheers,
> Viktor

-- 
Daimrod/Greg

[-- Attachment #2: Type: application/pgp-signature, Size: 835 bytes --]

Re: Wish: switch active time-stamps to inactive when CANCELED

[Karl Voit]

* Viktor Rosenfeld <listuser36@gmail.com> wrote:
> Hi,

Hi!

> Karl Voit wrote:
>
>> But still: there are two steps I have to make: (a) mark as canceled
>> because I want this "CANCELED" string attached and (b) add the
>> ARCHIVE tag.
>
> You can set the ARCHIVE tag when switching the TODO state automatically
> using org-todo-state-tags-triggers.
>
> For example (untestet):
>
> #+BEGIN_SRC emacs-lisp
> (setq org-todo-state-tags-triggers '(("CANCELLED" ("ARCHIVE" . t))))
> #+END_SRC

Well, I could not find any reason why I do not want this happen in
the same time. So, thanks for this trick - it solves indeed my issue
in a sleek way :-) 

-- 
Karl Voit

Re: Wish: switch active time-stamps to inactive when CANCELED

[Bastien]

Hi Viktor,

Viktor Rosenfeld <listuser36@gmail.com> writes:

> #+BEGIN_SRC emacs-lisp
> (setq org-todo-state-tags-triggers '(("CANCELLED" ("ARCHIVE" . t))))
> #+END_SRC

I shared this as a new FAQ, thanks!

-- 
 Bastien

Re: Wish: switch active time-stamps to inactive when CANCELED

[Viktor Rosenfeld]

Hi Bastien,

Bastien wrote:

> Viktor Rosenfeld <listuser36@gmail.com> writes:
> 
> > #+BEGIN_SRC emacs-lisp
> > (setq org-todo-state-tags-triggers '(("CANCELLED" ("ARCHIVE" . t))))
> > #+END_SRC
> 
> I shared this as a new FAQ, thanks!

The credit should go to Bernt Hansen. I found it in his org-mode setup.

Cheers,
Viktor

Re: Wish: switch active time-stamps to inactive when CANCELED

[Bastien]

Hi Viktor,

Viktor Rosenfeld <listuser36@gmail.com> writes:

> The credit should go to Bernt Hansen. I found it in his org-mode
> setup.

Fixed, thanks!

-- 
 Bastien

Re: Wish: switch active time-stamps to inactive when CANCELED

[Samuel Wales]

On 1/18/13, Karl Voit <devnull@karl-voit.at> wrote:
> Therefore, I am thinking that canceled events should turn their
> active time-stamps into inactive with an Org-mode option. This way,
> I can recall them in my agenda by typing ']' (show inactive
> time-stamps) but they do not interfere with my daily workflow.

Is an alternative to modify ] so that it will include active timestamps?

http://www.mail-archive.com/emacs-orgmode@gnu.org/msg57069.html

-- 
The Kafka Pandemic: http://thekafkapandemic.blogspot.com

The disease DOES progress.  MANY people have died from it.  ANYBODY
can get it.  There is no hope without action.

Re: Wish: switch active time-stamps to inactive when CANCELED

[Bastien]

Hi Karl,

Karl Voit <devnull@Karl-Voit.at> writes:

> I tend to keep events that I do not attend in my calendar for
> archiving reasons. But I switch their status to CANCELED.
> Unfortunately, those events still appear on my Google calendar and
> so forth. I get my usual reminders and it messes up my free/busy
> list.
>
> Therefore, I am thinking that canceled events should turn their
> active time-stamps into inactive with an Org-mode option. This way,
> I can recall them in my agenda by typing ']' (show inactive
> time-stamps) but they do not interfere with my daily workflow.
>
> What do you think of such an option for Org-mode?

I think this is what `org-after-todo-state-change-hook' is for :)

-- 
 Bastien

comma as decimal point: bug?

[Rainer Thiel]

I have seen that as of v. 7.9, org tables are supposed to accept
commas as decimal points which is useful if you set up tables to be
used in in files where continental conventions are expected to be
followed.

But I cannot see this really works.  I should have expected that the
following two tables should provide the same result:

-->
| Items  | Price |
|--------+-------|
| Item 1 |  15.3 |
| Item 2 |   2.3 |
|--------+-------|
| Sum    |  17.6 |
#+TBLFM: @4$2=vsum(@II..@I)

| Items  | Price |
|--------+-------|
| Item 1 |  15,3 |
| Item 2 |   2,3 |
|--------+-------|
| Sum    |    23 |
#+TBLFM: @4$2=vsum(@II..@I)
<--

As you can see, the outcome is quite different, and wrong when commas
are used instead of decimal points.

The value of org-table-number-regexp should be the standard value in org 7.9:

-->
org-table-number-regexp is a variable defined in `org-table.el'.
Its value is
"^\\([<>]?[-+^.,0-9]*[0-9][-+^.,0-9eEdDx()%:]*\\|\\(0[xX]\\)[0-9a-fA-F]+\\|nan\\)$"
<--

Am I doing wrong something here, or is this a bug in org 7.9?

Best regards, and many thanks in advance

Rainer
-- 
Prof. Dr. Rainer Thiel
Institut für Altertumswissenschaften
07737 Jena, Germany (EU)
r.thiel@uni-jena.de

Re: comma as decimal point: bug?

[Nicolas Goaziou]

Hello,

Rainer Thiel <r.thiel@uni-jena.de> writes:

> I have seen that as of v. 7.9, org tables are supposed to accept
> commas as decimal points which is useful if you set up tables to be
> used in in files where continental conventions are expected to be
> followed.

Actually, this more subtle. Tables recognize numbers with a comma as the
decimal mark. But it only means such numbers will be properly aligned to
the right. Internally, calculations still require dot separator.

Hope it is clearer now.


Regards,

-- 
Nicolas Goaziou

Re: comma as decimal point: bug?

[Bastien]

Nicolas Goaziou <n.goaziou@gmail.com> writes:

> But it only means such numbers will be properly aligned to
> the right. Internally, calculations still require dot separator.

Mhh.. but many users might expect the calculation to happen just 
because those strings are aligned as numbers.

The name of the variable is somewhat confusing here.

The default is a bit more rigid, but at least, it is not confusing,
and the variable (possibly renamed) can still be edited.

-- 
 Bastien

Re: comma as decimal point: bug?

[Nicolas Goaziou]

Hello,

Bastien <bzg@altern.org> writes:

> Mhh.. but many users might expect the calculation to happen just 
> because those strings are aligned as numbers.

">3.5" will also be recognized and aligned as a number, but you can't
expect calculations to happen on it either.


Regards,

-- 
Nicolas Goaziou

Re: comma as decimal point: bug?

[Bastien]

Hi all,

instead of having the comma as a decimal mark in all the regexps 
offered by the customization interface of `org-table-number-regexp',
I added the comma for a new choice, which has the most liberal
definition for what is a number.

Using this regexp as the default will confuse the users: if they see
"4,3" is aligned as a number, they will expect it to be treated as a
number.

This is not the case for ">3.5": even when such string is aligned as a
number, the user can clearly see by herself that this cannot be part of
an arithmetic operation.

I think the new option offered solves this issue.

Best,

-- 
 Bastien

Re: comma as decimal point: bug?

[Karl Voit]

* Nicolas Goaziou <n.goaziou@gmail.com> wrote:
>
> Rainer Thiel <r.thiel@uni-jena.de> writes:
>
>> I have seen that as of v. 7.9, org tables are supposed to accept
>> commas as decimal points which is useful if you set up tables to be
>> used in in files where continental conventions are expected to be
>> followed.
>
> Actually, this more subtle. Tables recognize numbers with a comma as the
> decimal mark. But it only means such numbers will be properly aligned to
> the right. Internally, calculations still require dot separator.

When I read the change-list of Org, I bet that there *will* be
problems because of this cosmetic change :-)

From user experience perspective, it is quite clear that «Org
recognizes numbers with commas» will be mixed up with «Org can
calculate with numbers containing commas». You can not make this
clear - even with renaming the settings variable.

Although I totally understand that numbers with commas should be
made possible (I am a German speaking guy) I guess there will be
even more issues in the future as long as cosmetics and
functionality is not aligned.

Just my 2 cents ...

-- 
Karl Voit

Re: comma as decimal point: bug?

[Bastien]

Hi Rainer,

Rainer Thiel <r.thiel@uni-jena.de> writes:

> I have seen that as of v. 7.9, org tables are supposed to accept
> commas as decimal points which is useful if you set up tables to be
> used in in files where continental conventions are expected to be
> followed.
>
> But I cannot see this really works.  

Well, the change was purely cosmetic, accepting commas by default 
will just change the way the table is aligned.  See this email by
Nicolas:

  http://article.gmane.org/gmane.emacs.orgmode/59134

But I guess it's confusing to align numbers as if they were numbers,
and to then tell the users "well, they are not really numbers."

I'm in favor of reverting this change and let the user change the
variable if he wants.

What do you - and Nicolas - think?

-- 
 Bastien

Re: comma as decimal point: bug?

[Rainer Thiel]

Hi Bastien, hi Nicolas,

2012/8/26 Bastien <bzg@altern.org>:
> But I guess it's confusing to align numbers as if they were numbers,
> and to then tell the users "well, they are not really numbers."

I myself indeed was confused, but now that I know it only concerns
aligning of numbers I have no problem living with it.  I had be
surprised by the supposed change, anyway, because Bastien (I think) in
earlier versions of the manual had most definitely excluded to allow
commas to be used as decimal points.

I am not a programmer and are probably unaware of the problems coming
up when allowing both dots and commas to be used as decimal points.
As a user, I can say it would be an advantage when exporting org-files
to LaTeX or ODT etc.

Many thanks again and best wishes

Rainer
-- 
Prof. Dr. Rainer Thiel
Institut für Altertumswissenschaften
07737 Jena, Germany (EU)
r.thiel@uni-jena.de

Re: comma as decimal point: bug?

[AW]

Am Montag, 27. August 2012, 11:08:59 schrieb Rainer Thiel:
> ...
> As a user, I can say it would be an advantage when exporting org-files
> to LaTeX or ODT etc.
> 
> Many thanks again and best wishes
> 
> Rainer

Exporting tables to LaTeX usually is not the last step, obviously you will 
compile the exported *.tex file into a *.pdf. Right?

So stick to a dot as the decimal point in *.org and in *.tex, but use the 
numprint package to get a table in your *.pdf with commas as decimal points.

We had that discussion here and William Henney gave an example in an email 
dated 22 April 2012. 

Regards,
Alexander

Replacement for org-make-link

[Karl Voit]

Hi!

http://orgmode.org/Changes.html tells me that org-make-link has been
deleted.

I am using following lines to add links to references I manage using
Org-mode: (from [1])

,----
| (defun org-ref-complete-link (&optional arg)
|   "Create a reference link using completion."
|   (let (file link)
|        (setq file (read-file-name "ref: " "~/archive/papers_from_web/"))
|        (let ((pwd (file-name-as-directory (expand-file-name ".")))
|          (pwd1 (file-name-as-directory (abbreviate-file-name
|                         (expand-file-name ".")))))
|          (setq file (string-replace "~/archive/papers_from_web/" "" file))
|          (setq file (string-replace pwd "" (string-replace pwd1 "" file)))
|          (setq file (string-replace ".bib" "" file))
|          (setq file (string-replace ".pdf" "" file))
|          (setq link (org-make-link "ref:" file)))
|     link))
`----

Can somebody with more knowledge than me suggest, what I should do
instead?

Thanks!


  1. http://draketo.de/light/english/free-software/custom-link-completion-org-mode-25-lines-emacs
-- 
Karl Voit

table: referencing row of other table

[Karl Voit]

Hi!

I want to use the column "average" of the first table to fill the
column "h1" in the second one.

#+TBLNAME: 2012-08-12vkmeasure
| tags/item |    m1 |    m2 |    m3 | average |
|-----------+-------+-------+-------+---------|
|         4 |  0.02 |  0.03 |  0.02 |    0.02 |
|         5 |  0.06 |  0.12 |  0.06 |    0.08 |
|         6 |  0.31 |  0.53 |  0.29 |    0.38 |
|         7 |  3.83 |  4.08 |  4.48 |    4.13 |
|         8 | 85.33 | 89.22 | 92.07 |   88.87 |
#+TBLFM: $5 = vmean($2..$4);%.2f

| tags/item | h1 [s] | h2 [s] |
|-----------+--------+--------|
|         4 |        |        |
|         5 |        |        |
|         6 |        |        |
|         7 |        |        |
|         8 |        |        |
#+TBLFM: <fillme>

I tried following references but failed so far:

@2$2..@6$2=subscr(remote(2012-08-12vkmeasure,@2$5..@6$5),$#) #-> 0.08 0.08 ...
@2$2=remote(2012-08-12vkmeasure,@2$5) -> works for one value
@2$2..@6$2=subscr(remote(2012-08-12vkmeasure,@2$5..@6$5),$#) #-> 0.08 0.08 ...
$2=subscr(remote(2012-08-12vkmeasure,@I$>..@II$>),$#) 

What did I do wrong?

How is the correct TBLFM in order to get 0.02 ... 88.87 into the
h1-column?

Thanks for your help!

-- 
Karl Voit

Plotting (with gnuplot) using dates timestamps

[Alan Schmitt]

Hello,

I'm trying to plot the following table, but the dates part is all wrong. Is there a way to tell gnuplot what the date format is?

#+tblname: data-table
| Date             |     HP |     HC |
|------------------+--------+--------|
| [2011-08-20 Sat] | 006815 | 008399 |
| [2011-08-29 Mon] | 006840 | 008438 |
| [2011-09-11 Sun] | 006946 | 008552 |
| [2011-12-11 Sun] | 007805 | 009603 |
| [2012-03-04 Sun] | 008800 | 010826 |
| [2012-03-11 Sun] | 008876 | 010930 |
| [2012-03-25 Sun] | 009015 | 011121 |

#+begin_src gnuplot :var data=data-table
reset

set title "Consommation Électrique"
set size ratio square

set xlabel "Date"

set yrange [6800:9100]
set ylabel "HP"
set ytics nomirror

set y2range [8000:12000]
set y2label "HC"
set y2tics nomirror

set style data points
plot data using 1:2 axis x1y1 title 'HP', \
     data using 1:3 axis x1y2 title 'HC'
#+end_src


Thanks,

Alan

Re: Plotting (with gnuplot) using dates timestamps

[Allen S. Rout]

On 03/26/2012 09:37 AM, Alan Schmitt wrote:
> Hello,
>
> I'm trying to plot the following table, but the dates part is all wrong. Is there a way to tell gnuplot what the date format is?
>

I've found gnuplot's date support to be torment.

You might want to investigate using R for these tasks:  R understands 
what a date is at a fundamental level, and doesn't give you any crap 
about it.

Better still, once you climb over the activation energy of changing 
graphing idioms, a huge vista of new capabilities are open to you.

- Allen S. Rout

Re: Plotting (with gnuplot) using dates timestamps

[Michael Brand]

Hi Alan

On Mon, Mar 26, 2012 at 15:37, Alan Schmitt
<alan.schmitt@polytechnique.org> wrote:
> I'm trying to plot the following table, but the dates part is all wrong. Is there a way to tell gnuplot what the date format is?
>
> #+tblname: data-table
> | Date             |     HP |     HC |
> |------------------+--------+--------|
> | [2011-08-20 Sat] | 006815 | 008399 |

Did you try this?
set timefmt  "[%Y-%m-%d"

Michael

Re: Plotting (with gnuplot) using dates timestamps

[Steven Buczkowski]

On Mon, 2012-03-26 at 15:37 +0200, Alan Schmitt wrote:
> Hello,
> 
> I'm trying to plot the following table, but the dates part is all wrong. Is there a way to tell gnuplot what the date format is?
> 
> #+tblname: data-table
> | Date             |     HP |     HC |
> |------------------+--------+--------|
> | [2011-08-20 Sat] | 006815 | 008399 |
> | [2011-08-29 Mon] | 006840 | 008438 |
> | [2011-09-11 Sun] | 006946 | 008552 |
> | [2011-12-11 Sun] | 007805 | 009603 |
> | [2012-03-04 Sun] | 008800 | 010826 |
> | [2012-03-11 Sun] | 008876 | 010930 |
> | [2012-03-25 Sun] | 009015 | 011121 |
> 

In straight gnuplot, I would do something like the following:

gnuplot> set xdata time
gnuplot> set timefmt '[%Y-%m-%d %a]'

then, plot. 

Steven


-- 
-----------------------------------
Steven Buczkowski
PhD Candidate - Atmospheric Physics
UMBC Physics Dept
1000 Hilltop Circle
Baltimore, MD 21250

Email: sbuczkowski@umbc.edu
Cell: 413-687-9102 (No texts. Email instead.)
2m/70cm: KB3RRJ (147.135+[156.7]/448.275-[156.7])
Smoke Signal: Two long, Two short. I'll send a rider.
-----------------------------------

Re: Plotting (with gnuplot) using dates timestamps

[Nick Dokos]

Steven Buczkowski <steven.buczkowski@gmail.com> wrote:

> On Mon, 2012-03-26 at 15:37 +0200, Alan Schmitt wrote:
> > Hello,
> > 
> > I'm trying to plot the following table, but the dates part is all wrong. Is there a way to tell gnuplot what the date format is?
> > 
> > #+tblname: data-table
> > | Date             |     HP |     HC |
> > |------------------+--------+--------|
> > | [2011-08-20 Sat] | 006815 | 008399 |
> > | [2011-08-29 Mon] | 006840 | 008438 |
> > | [2011-09-11 Sun] | 006946 | 008552 |
> > | [2011-12-11 Sun] | 007805 | 009603 |
> > | [2012-03-04 Sun] | 008800 | 010826 |
> > | [2012-03-11 Sun] | 008876 | 010930 |
> > | [2012-03-25 Sun] | 009015 | 011121 |
> > 
> 
> In straight gnuplot, I would do something like the following:
> 
> gnuplot> set xdata time
> gnuplot> set timefmt '[%Y-%m-%d %a]'
> 

The "set xdata time" line is indeed what makes the difference:
without it, gnuplot doesn't know that this is a time series.
But there are some additional details: when babel prepares the
data, the data file that will be fed to gnuplot ends up like this
(not sure this is documented though):

,----
| 2011-08-20-00:00:00	6815	8399
| 2011-08-29-00:00:00	6840	8438
| 2011-09-11-00:00:00	6946	8552
| 2011-12-11-00:00:00	7805	9603
| 2012-03-04-00:00:00	8800	10826
| 2012-03-11-00:00:00	8876	10930
| 2012-03-25-00:00:00	9015	11121
`----

so the script should look something like this:

--8<---------------cut here---------------start------------->8---
reset

set xdata time
set timefmt "%Y-%m-%d-%H:%M:%S"
set format x "%Y-%m-%d"
set xrange ["2011-08-01":"2012-04-01"]

set title "Consumption"
set size ratio square

set xlabel "Date"

set yrange [6800:9100]
set ylabel "HP"
set ytics nomirror

set y2range [8000:12000]
set y2label "HC"
set y2tics nomirror

set style data points
plot data using 1:2 axis x1y1 title 'HP', \
     data using 1:3 axis x1y2 title 'HC'
--8<---------------cut here---------------end--------------->8---


BTW, going into gnuplot and saying "help time/date" helps :-)
I have trouble navigating the gnuplot help system, but I think
it is complete.

Nick

Re: Plotting (with gnuplot) using dates timestamps

[Alan Schmitt]

On 26 mars 2012, at 16:37, Nick Dokos wrote:

> Steven Buczkowski <steven.buczkowski@gmail.com> wrote:
> 
>> On Mon, 2012-03-26 at 15:37 +0200, Alan Schmitt wrote:
>>> Hello,
>>> 
>>> I'm trying to plot the following table, but the dates part is all wrong. Is there a way to tell gnuplot what the date format is?
>>> 
>>> #+tblname: data-table
>>> | Date             |     HP |     HC |
>>> |------------------+--------+--------|
>>> | [2011-08-20 Sat] | 006815 | 008399 |
>>> | [2011-08-29 Mon] | 006840 | 008438 |
>>> | [2011-09-11 Sun] | 006946 | 008552 |
>>> | [2011-12-11 Sun] | 007805 | 009603 |
>>> | [2012-03-04 Sun] | 008800 | 010826 |
>>> | [2012-03-11 Sun] | 008876 | 010930 |
>>> | [2012-03-25 Sun] | 009015 | 011121 |
>>> 
>> 
>> In straight gnuplot, I would do something like the following:
>> 
>> gnuplot> set xdata time
>> gnuplot> set timefmt '[%Y-%m-%d %a]'
>> 
> 
> The "set xdata time" line is indeed what makes the difference:
> without it, gnuplot doesn't know that this is a time series.
> But there are some additional details: when babel prepares the
> data, the data file that will be fed to gnuplot ends up like this
> (not sure this is documented though):
> 
> ,----
> | 2011-08-20-00:00:00	6815	8399
> | 2011-08-29-00:00:00	6840	8438
> | 2011-09-11-00:00:00	6946	8552
> | 2011-12-11-00:00:00	7805	9603
> | 2012-03-04-00:00:00	8800	10826
> | 2012-03-11-00:00:00	8876	10930
> | 2012-03-25-00:00:00	9015	11121
> `----

Thanks a lot for all the suggestions. The part I was missing was the preprocessing from babel. Using simply

set xdata time
set timefmt "%Y-%m-%d"

did the trick.

And yes: the gnuplot built-in help seems very complete.

Thanks again,

Alan

Re: Plotting (with gnuplot) using dates timestamps

[Karl Voit]

* Alan Schmitt <alan.schmitt@polytechnique.org> wrote:
>
> Thanks a lot for all the suggestions. The part I was missing was the preprocessing from babel. Using simply
>
> set xdata time
> set timefmt "%Y-%m-%d"
>
> did the trick.

I just wanted to check this out but I am not able to produce the
plot :-( I just get the coordinate system but with an empty plane
(no points).

Can you please post the while working example (or correct my error)?

Thanks!


#+tblname: data-table
| Date             |     HP |     HC |
|------------------+--------+--------|
| [2011-08-20 Sat] | 006815 | 008399 |
| [2011-08-29 Mon] | 006840 | 008438 |
| [2011-09-11 Sun] | 006946 | 008552 |
| [2011-12-11 Sun] | 007805 | 009603 |
| [2012-03-04 Sun] | 008800 | 010826 |
| [2012-03-11 Sun] | 008876 | 010930 |
| [2012-03-25 Sun] | 009015 | 011121 |

#+begin_src gnuplot :var data=data-table
reset

set title "Consommation Électrique"
set size ratio square

set xlabel "Date"
set xdata time
set timefmt "[%Y-%m-%d %a]"
set xrange ["[2011-08-01 foo]":"[2012-04-01 bar]"]

set yrange [6800:9100]
set ylabel "HP"
set ytics nomirror

set y2range [8000:12000]
set y2label "HC"
set y2tics nomirror

set style data points
plot data using 1:2 axis x1y1 title 'HP', \
     data using 1:3 axis x1y2 title 'HC'
#+end_src


-- 
Karl Voit

Hallo, mein Name ist Web 2.0. Gib mir dein Adressbuch und lass mich Emails lesen

Re: Plotting (with gnuplot) using dates timestamps

[Alan Schmitt]

On 26 mars 2012, at 17:48, Karl Voit wrote:

> * Alan Schmitt <alan.schmitt@polytechnique.org> wrote:
>> 
>> Thanks a lot for all the suggestions. The part I was missing was the preprocessing from babel. Using simply
>> 
>> set xdata time
>> set timefmt "%Y-%m-%d"
>> 
>> did the trick.
> 
> I just wanted to check this out but I am not able to produce the
> plot :-( I just get the coordinate system but with an empty plane
> (no points).
> 
> Can you please post the while working example (or correct my error)?

Change this

> set timefmt "[%Y-%m-%d %a]"

to this

set timefmt "%Y-%m-%d"

I also don't specify the range.

> set xrange ["[2011-08-01 foo]":"[2012-04-01 bar]"]

Here is the full version.

#+tblname: data-table
| Date             |     HP |     HC |
|------------------+--------+--------|
| [2011-08-20 Sat] | 006815 | 008399 |
| [2011-08-29 Mon] | 006840 | 008438 |
| [2011-09-11 Sun] | 006946 | 008552 |
| [2011-12-11 Sun] | 007805 | 009603 |
| [2012-03-04 Sun] | 008800 | 010826 |
| [2012-03-11 Sun] | 008876 | 010930 |
| [2012-03-25 Sun] | 009015 | 011121 |

#+begin_src gnuplot :var data=data-table
reset

set title "Consommation Électrique"
set size ratio square

set xlabel "Date"
set xdata time
set timefmt "%Y-%m-%d"
set format x "%Y-%m"
set xtics nomirror

set ylabel "HP"
set ytics nomirror

set y2label "HC"
set y2tics nomirror

set style data points
plot data using 1:2 with linespoints axis x1y1 title 'HP', \
    data using 1:3 with linespoints axis x1y2 title 'HC'
#+end_src

Alan

Re: Plotting (with gnuplot) using dates timestamps

[Karl Voit]

* Alan Schmitt <alan.schmitt@polytechnique.org> wrote:
> On 26 mars 2012, at 17:48, Karl Voit wrote:
>
>> Can you please post the while working example (or correct my error)?
>
> Change this
>> set timefmt "[%Y-%m-%d %a]"
> to this
> set timefmt "%Y-%m-%d"
>
> I also don't specify the range.
>> set xrange ["[2011-08-01 foo]":"[2012-04-01 bar]"]

Thanks!

I now understand and it works :-)

-- 
Karl Voit

Re: Plotting (with gnuplot) using dates timestamps

[Ian Barton]

On 26/03/12 18:06, Karl Voit wrote:
> * Alan Schmitt<alan.schmitt@polytechnique.org>  wrote:
>> On 26 mars 2012, at 17:48, Karl Voit wrote:
>>
>>> Can you please post the while working example (or correct my error)?
>>
>> Change this
>>> set timefmt "[%Y-%m-%d %a]"
>> to this
>> set timefmt "%Y-%m-%d"
>>
>> I also don't specify the range.
>>> set xrange ["[2011-08-01 foo]":"[2012-04-01 bar]"]
>
> Thanks!
>
> I now understand and it works :-)
>

Haivng played with gnuplot recently, I would suggest that your first 
statement should always be "reset". Otherwise you may find settings you 
have been trying out persist over different runs of your babel code. 
This can be very confusing and may lead to premature hair loss.

Ian.

Re: Plotting (with gnuplot) using dates timestamps

[Karl Voit]

* Ian Barton <lists@wilkesley.net> wrote:
>
> Haivng played with gnuplot recently, I would suggest that your first 
> statement should always be "reset". Otherwise you may find settings you 
> have been trying out persist over different runs of your babel code. 

Great tipp, thanks!

> This can be very confusing and may lead to premature hair loss.

Since you actually can not see me: way too late :-)

-- 
Karl Voit

Re: Plotting (with gnuplot) using dates timestamps

[Mueen Nawaz]

Alan Schmitt <alan.schmitt@polytechnique.org> writes:

> Hello,
>
> I'm trying to plot the following table, but the dates part is all
> wrong. Is there a way to tell gnuplot what the date format is?
>
> #+tblname: data-table
> | Date             |     HP |     HC |
> |------------------+--------+--------|
> | [2011-08-20 Sat] | 006815 | 008399 |
> | [2011-08-29 Mon] | 006840 | 008438 |
> | [2011-09-11 Sun] | 006946 | 008552 |
> | [2011-12-11 Sun] | 007805 | 009603 |
> | [2012-03-04 Sun] | 008800 | 010826 |
> | [2012-03-11 Sun] | 008876 | 010930 |
> | [2012-03-25 Sun] | 009015 | 011121 |


Although you solved the problem, another solution for headaches like
these is simply to preprocess the table in something like Perl/Python
(within your org file using Babel) to get the desired string
format. Then pass the resulting array/list to gnuplot.

The solution you were provided is probably the way to go, but keep this
strategy in mind - sometimes it's just not worth the effort to solve
everything in one language.

[bug] org-agenda-write does not handle date stamps without day of week

[Karl Voit]

Hi!

I verified the behavior below with the most current git version of
Org-mode.

Whenever I call my function to export the current agenda, it
«converts» events that do have a beginning and end time to events
that span the whole day (without begin and end times at all).

,----[ my LISP command to export ]
| (defun vk-export-agenda()
|   "Exports monthly Org-mode agenda to agenda.ics file"
|   (interactive)
|   (org-agenda-list nil nil 60)
|   (org-agenda-write "~/org/agenda.ics")
| )
`----

I traced down the problem to this:

* <2012-03-05 Mon 08:00-09:00> works fine
* <2012-03-05 Mon 8:00-9:00> works fine too

* <2012-03-05 08:00-09:00> Wrong: ends up as full day event

Unfortunately, I do have certain mechanisms that generate parts of
my Org-mode files where I skip the day of the week. Org-mode in
general is able to handle date stamps without day of week pretty
well. But whenever I export it to ics (in order to send it to
Google), I end up with a messed up online agenda.

I guess that org-agenda-write is the one to blame. It would be very
handy to me if somebody could fix that issue.

Let me know, if you need any further information!

Thank you very much!

-- 
Karl Voit

Re: [bug] org-agenda-write does not handle date stamps without day of week

[Karl Voit]

Since no followup reaction occured: is this a but that requires much
work? Or did my email just got lost in overfull inboxes? :-)

* Karl Voit <devnull@Karl-Voit.at> wrote:
>
> I verified the behavior below with the most current git version of
> Org-mode.
>
> Whenever I call my function to export the current agenda, it
> «converts» events that do have a beginning and end time to events
> that span the whole day (without begin and end times at all).
>
> ,----[ my LISP command to export ]
>| (defun vk-export-agenda()
>|   "Exports monthly Org-mode agenda to agenda.ics file"
>|   (interactive)
>|   (org-agenda-list nil nil 60)
>|   (org-agenda-write "~/org/agenda.ics")
>| )
> `----
>
> I traced down the problem to this:
>
> * <2012-03-05 Mon 08:00-09:00> works fine
> * <2012-03-05 Mon 8:00-9:00> works fine too
>
> * <2012-03-05 08:00-09:00> Wrong: ends up as full day event
>
> Unfortunately, I do have certain mechanisms that generate parts of
> my Org-mode files where I skip the day of the week. Org-mode in
> general is able to handle date stamps without day of week pretty
> well. But whenever I export it to ics (in order to send it to
> Google), I end up with a messed up online agenda.
>
> I guess that org-agenda-write is the one to blame. It would be very
> handy to me if somebody could fix that issue.
>
> Let me know, if you need any further information!
>
> Thank you very much!

-- 
Karl Voit

Re: [bug] org-agenda-write does not handle date stamps without day of week

[Simon Thum]

Hi Karl,

why don't you add something like an "xx" day which will probably please 
org-mode - in my tests org ignored the weekday. I concluded it's there 
for viewing convenience, and editing (C-.) always got me the right DOW.

I don't use org2ical, however.

Cheers,

Simon


On 03/05/2012 03:44 PM, Karl Voit wrote:
> Hi!
>
> I verified the behavior below with the most current git version of
> Org-mode.
>
> Whenever I call my function to export the current agenda, it
> «converts» events that do have a beginning and end time to events
> that span the whole day (without begin and end times at all).
>
> ,----[ my LISP command to export ]
> | (defun vk-export-agenda()
> |   "Exports monthly Org-mode agenda to agenda.ics file"
> |   (interactive)
> |   (org-agenda-list nil nil 60)
> |   (org-agenda-write "~/org/agenda.ics")
> | )
> `----
>
> I traced down the problem to this:
>
> *<2012-03-05 Mon 08:00-09:00>  works fine
> *<2012-03-05 Mon 8:00-9:00>  works fine too
>
> *<2012-03-05 08:00-09:00>  Wrong: ends up as full day event
>
> Unfortunately, I do have certain mechanisms that generate parts of
> my Org-mode files where I skip the day of the week. Org-mode in
> general is able to handle date stamps without day of week pretty
> well. But whenever I export it to ics (in order to send it to
> Google), I end up with a messed up online agenda.
>
> I guess that org-agenda-write is the one to blame. It would be very
> handy to me if somebody could fix that issue.
>
> Let me know, if you need any further information!
>
> Thank you very much!
>

Re: [bug] org-agenda-write does not handle date stamps without day of week

[Carsten Dominik]

On 17.3.2012, at 15:04, Simon Thum wrote:

> Hi Karl,
> 
> why don't you add something like an "xx" day which will probably please org-mode - in my tests org ignored the weekday.

Yes, the weekday is only expected, but never used, so writing "xxx"
would work.  However, see my other mail for another proposal
along Nick's lines of argument.

- Carsten


> I concluded it's there for viewing convenience, and editing (C-.) always got me the right DOW.
> 
> I don't use org2ical, however.
> 
> Cheers,
> 
> Simon
> 
> 
> On 03/05/2012 03:44 PM, Karl Voit wrote:
>> Hi!
>> 
>> I verified the behavior below with the most current git version of
>> Org-mode.
>> 
>> Whenever I call my function to export the current agenda, it
>> «converts» events that do have a beginning and end time to events
>> that span the whole day (without begin and end times at all).
>> 
>> ,----[ my LISP command to export ]
>> | (defun vk-export-agenda()
>> |   "Exports monthly Org-mode agenda to agenda.ics file"
>> |   (interactive)
>> |   (org-agenda-list nil nil 60)
>> |   (org-agenda-write "~/org/agenda.ics")
>> | )
>> `----
>> 
>> I traced down the problem to this:
>> 
>> *<2012-03-05 Mon 08:00-09:00>  works fine
>> *<2012-03-05 Mon 8:00-9:00>  works fine too
>> 
>> *<2012-03-05 08:00-09:00>  Wrong: ends up as full day event
>> 
>> Unfortunately, I do have certain mechanisms that generate parts of
>> my Org-mode files where I skip the day of the week. Org-mode in
>> general is able to handle date stamps without day of week pretty
>> well. But whenever I export it to ics (in order to send it to
>> Google), I end up with a messed up online agenda.
>> 
>> I guess that org-agenda-write is the one to blame. It would be very
>> handy to me if somebody could fix that issue.
>> 
>> Let me know, if you need any further information!
>> 
>> Thank you very much!
>> 
> 
> 

Recurring events with exceptions

[Karl Voit]

Hi!

I am into a process to write a convert tool from my old calendar
software[1] to Org-mode.

Now I do have to define something like »this event is recurring each
week on Wednesday except 2011-10-26 and 2011-11-30«.

I already know that complex things have to be done using sexp
entries[2] but this does not seem to be possible with sexp either.

Before I do have to develop a method that generates multiple
distinct events for each recurring definition: is there another way
to achieve this?

Thanks!

  1. jPilot/DateBK6/PalmOS
  2. http://www.gnu.org/software/emacs/manual/html_node/emacs/Sexp-Diary-Entries.html
-- 
Karl Voit

Re: Recurring events with exceptions

[Brian Wightman]

(and (your-sexp-here) (not (except-dates-here)))

On Tue, Oct 18, 2011 at 11:52 AM, Karl Voit <devnull@karl-voit.at> wrote:
> Hi!
>
> I am into a process to write a convert tool from my old calendar
> software[1] to Org-mode.
>
> Now I do have to define something like »this event is recurring each
> week on Wednesday except 2011-10-26 and 2011-11-30«.
>
> I already know that complex things have to be done using sexp
> entries[2] but this does not seem to be possible with sexp either.
>
> Before I do have to develop a method that generates multiple
> distinct events for each recurring definition: is there another way
> to achieve this?
>
> Thanks!
>
>  1. jPilot/DateBK6/PalmOS
>  2. http://www.gnu.org/software/emacs/manual/html_node/emacs/Sexp-Diary-Entries.html
> --
> Karl Voit
>
>
>

Re: Recurring events with exceptions

[Karl Voit]

* Brian Wightman <brian@wightmanfam.org> wrote:
> (and (your-sexp-here) (not (except-dates-here)))

Cool, thanks!

Now I still have to think about which method is easier to implement
:-)

-- 
Karl Voit

Re: Recurring events with exceptions

[Skip Collins]

> (and (your-sexp-here) (not (except-dates-here)))

Pardon a slightly off-topic rant. I have yet to find calendar software
(org included) that handles repeating appointments with the kind of
flexibility that would make them really useful. First, typical
real-world repeating appointments do not follow a rigid pattern such
as "visit the gym every monday from 7pm to 9pm." I know that I will
not be following that pattern 50 years from now. But the default
repeat pattern in most software does not include an end date.
Infinitely repeating appointments are just stupid. Second, the whole
concept of an exception to a repeating appointment is broken. It
presumes that the pattern is the norm and the exception is, well,
exceptional. None of my weekly meetings actually meets every week in a
given year. The code above illustrates the problem. What if I just
want to move one instance by a half-hour? The difference between
deleting an instance and changing or adding an instance is hard to
capture when basing everything off of a rigid pattern.

In my opinion, a better framework for dealing with repeating
appointments would be to view them as a list of individual instances
that can be managed as a collection. When creating a repeating
appointment, the user should be able to generate a list that follows
some pattern (e.g. every monday at 7pm), with a mandatory end date.
The user should also be able to easily add or subtract arbitrary
instances to or from the list. Actions can be performed on all
instances in a collection or on individual instances. The collection
of instances could be defined by a shared UID or tag. But each
instance should be flexible enough to have its own title, date, time,
duration, invitee list, location, etc. It should also be possible to
separate (or copy) individual instances from the repeat collection.
Using arbitrarily complex sexp functions to achieve reasonably
flexible repeating appointments seems like overkill when a properly
structured and managed list can accomplish the desired effect.

Of course the devil is in the implementation details.

Re: Recurring events with exceptions

[Tassilo Horn]

Skip Collins <skip.collins@gmail.com> writes:

Hi Skip,

> [a slightly off-topic rant]

You might find some luck with applications that follow the iCalendar
standard.  There, a re-occuring event may have an end date.  And when
you move ("this week's meeting is 30 minutes later") or delete some
occurence ("no meeting on christmas"), it will add that as an exception
to the re-occuring event definition and clone that as a new once-only
event in case of a move (probably keeping some connection with it).

[But I think there's no concept of re-occuring exceptions, like "there
will never be any meetings at Christmas"...]

That said, this is pretty complex and the apps I've used before org-mode
(KOrganizer, Evolution) managed to mess things up in certain cases
without even being visible to me at first.  I just noticed some time
later that some appointments were simply wrong or totally missing.

I've learned my lessons: now I accept infinitely repeating entries or
switch to sexp entries if really needed.  But at least I know that
nothing I do will somehow corrupt my calendar, because *I* am the master
of timestamps, not some GUI that DTRT only in 90% of the cases.

Bye,
Tassilo

Re: Recurring events with exceptions

[Eric S Fraga]

Skip Collins <skip.collins@gmail.com> writes:

>> (and (your-sexp-here) (not (except-dates-here)))
>
> Pardon a slightly off-topic rant. I have yet to find calendar software
> (org included) that handles repeating appointments with the kind of
> flexibility that would make them really useful. First, typical

[...]

Have you looked at the following?

,----[ C-h f org-clone-subtree-with-time-shift RET ]
| org-clone-subtree-with-time-shift is an interactive compiled Lisp
| function in `org.el'.
| 
| (org-clone-subtree-with-time-shift N &optional SHIFT)
| 
| Clone the task (subtree) at point N times.
| The clones will be inserted as siblings.
| 
| In interactive use, the user will be prompted for the number of
| clones to be produced, and for a time SHIFT, which may be a
| repeater as used in time stamps, for example `+3d'.
| 
| When a valid repeater is given and the entry contains any time
| stamps, the clones will become a sequence in time, with time
| stamps in the subtree shifted for each clone produced.  If SHIFT
| is nil or the empty string, time stamps will be left alone.  The
| ID property of the original subtree is removed.
| 
| If the original subtree did contain time stamps with a repeater,
| the following will happen:
| - the repeater will be removed in each clone
| - an additional clone will be produced, with the current, unshifted
|   date(s) in the entry.
| - the original entry will be placed *after* all the clones, with
|   repeater intact.
| - the start days in the repeater in the original entry will be shifted
|   to past the last clone.
| I this way you can spell out a number of instances of a repeating task,
| and still retain the repeater to cover future instances of the task.
| 
| [back]
`----

I find that this exactly what you are describing: I define an entry,
clone it (with weekly shift, for instance) and then delete the
exceptions and maybe add a few extras.  If any weekly instance has to
change, I simply change it!  Works pretty well and handles all the cases
you mentioned.


-- 
: Eric S Fraga (GnuPG: 0xC89193D8FFFCF67D) in Emacs 24.0.90.1
: using Org-mode version 7.7 (release_7.7.381.g05ea.dirty)

Re: Recurring events with exceptions

[Skip Collins]

> ,----[ C-h f org-clone-subtree-with-time-shift RET ]
>
> I find that this exactly what you are describing: I define an entry,
> clone it (with weekly shift, for instance) and then delete the
> exceptions and maybe add a few extras.  If any weekly instance has to
> change, I simply change it!  Works pretty well and handles all the cases
> you mentioned.

There is one important feature lacking in the use of time-shifted
clones as a way to generate repeating events: there is nothing to bind
individual instances into a collection. A common EVENTID property
would allow for the development of tools for handling a whole series
as a group to do things like deleting, shifting, and making other
changes en masse. It seems like a simple thing to add.

Flexibility like this is why I like working in org. I would move my
calendar to org if I were not stuck with outlook/exchange at work. I'm
still waiting for someone to create the ultimate org/exchange sync
solution.

Re: Recurring events with exceptions

[Eric S Fraga]

Skip Collins <skip.collins@gmail.com> writes:

>> ,----[ C-h f org-clone-subtree-with-time-shift RET ]
>>
>
>> I find that this exactly what you are describing: I define an entry,
>> clone it (with weekly shift, for instance) and then delete the
>> exceptions and maybe add a few extras.  If any weekly instance has to
>> change, I simply change it!  Works pretty well and handles all the cases
>> you mentioned.
>
> There is one important feature lacking in the use of time-shifted
> clones as a way to generate repeating events: there is nothing to bind
> individual instances into a collection. A common EVENTID property
> would allow for the development of tools for handling a whole series
> as a group to do things like deleting, shifting, and making other
> changes en masse. It seems like a simple thing to add.

I'm probably missing a key point but can you not simply put a property
in a higher level heading that encloses all the cloned entries?

> Flexibility like this is why I like working in org. I would move my
> calendar to org if I were not stuck with outlook/exchange at work. I'm
> still waiting for someone to create the ultimate org/exchange sync
> solution.

Yes, I have been informed that we are moving to Outlook/exchange in the
new year, something I am dreading...  Matthieu Lemerre, on this list
back in late June (Message-ID: <871uygg1fs.fsf@free.fr>), posted a
(partial?) solution that helped in this regard.  I've not tried that
solution as I don't have to use Outlook yet.  Search the mailing list.

-- 
: Eric S Fraga (GnuPG: 0xC89193D8FFFCF67D) in Emacs 24.0.90.1
: using Org-mode version 7.7 (release_7.7.381.g05ea.dirty)

Re: Recurring events with exceptions

[Skip Collins]

On Wed, Oct 19, 2011 at 6:02 AM, Eric S Fraga <e.fraga@ucl.ac.uk> wrote:
> Yes, I have been informed that we are moving to Outlook/exchange in the
> new year, something I am dreading...  Matthieu Lemerre, on this list
> back in late June (Message-ID: <871uygg1fs.fsf@free.fr>), posted a
> (partial?) solution that helped in this regard.  I've not tried that
> solution as I don't have to use Outlook yet.  Search the mailing list.

Lemerre's approach only syncs with the Outlook client, not the
Exchange server. He suggests using RFC-2446 (iCalendar) as a basis for
a more general solution. I think that trying to establish a smoothly
syncing connection between org and exchange by using iCalendar is a
recipe for frustration. This is mostly because Exchange's
implementation of the standard is lacking.

It might be more fruitful to pursue an org interface to Exchange Web
Services, which is favored and privileged by microsoft. A hypothetical
org-ews.el would pass xml messages that look something like:

<?xml version="1.0" encoding="utf-8"?>
<soap:Envelope xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
               xmlns:m="http://schemas.microsoft.com/exchange/services/2006/messages"
               xmlns:t="http://schemas.microsoft.com/exchange/services/2006/types"
               xmlns:soap="http://schemas.xmlsoap.org/soap/envelope/">
   <soap:Header>
      <t:RequestServerVersion Version="Exchange2010" />
   </soap:Header>
   <soap:Body>
      <m:CreateItem SendMeetingInvitations="SendToNone">
         <m:Items>
            <t:CalendarItem>
               <t:Subject>Dentist Appointment</t:Subject>
               <t:Body BodyType="Text">The appointment is with Dr.
Smith.</t:Body>
               <t:Start>2009-03-02T17:00:00Z</t:Start>
               <t:End>2009-03-02T19:00:00Z</t:End>
            </t:CalendarItem>
         </m:Items>
      </m:CreateItem>
   </soap:Body>
</soap:Envelope>

Perhaps we could entice Jambunathan to take this on ;-)
His experience getting org to speak xml as a way to interact with
other systems would be invaluable.

Sync with Outlook/Exchange (was: Recurring events with exceptions)

[Karl Voit]

* Skip Collins <skip.collins@gmail.com> wrote:
> On Wed, Oct 19, 2011 at 6:02 AM, Eric S Fraga <e.fraga@ucl.ac.uk> wrote:
>> Yes, I have been informed that we are moving to Outlook/exchange in the
>> new year, something I am dreading...
>
> Lemerre's approach only syncs with the Outlook client, not the
> Exchange server. He suggests using RFC-2446 (iCalendar) as a basis for
> a more general solution. I think that trying to establish a smoothly
> syncing connection between org and exchange by using iCalendar is a
> recipe for frustration. This is mostly because Exchange's
> implementation of the standard is lacking.

Just a side note:

In my opinion, this is a *very important issue* for Org-mode.
Personally I know several people who would love to use
Emacs/Org-mode but are forced to use Outlook because of Exchange in
their office.

Whenever I present Org-mode to a variety of people, the question of
Outlook/Exchange support is one of the first being asked.

For now I am in the lucky position of no dependency to Exchange but I
guess this can change in the future.


PS: Please adopt the subject line to changing subject - especially
for important ones like here. Thanks! :-)

-- 
Karl Voit

Re: Recurring events with exceptions

[Eric S Fraga]

Skip Collins <skip.collins@gmail.com> writes:


[...]

> Lemerre's approach only syncs with the Outlook client, not the
> Exchange server. He suggests using RFC-2446 (iCalendar) as a basis for
> a more general solution. I think that trying to establish a smoothly
> syncing connection between org and exchange by using iCalendar is a
> recipe for frustration. This is mostly because Exchange's
> implementation of the standard is lacking.
>
> It might be more fruitful to pursue an org interface to Exchange Web
> Services, which is favored and privileged by microsoft. A hypothetical
> org-ews.el would pass xml messages that look something like:
>
> <?xml version="1.0" encoding="utf-8"?>
> <soap:Envelope xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"

[...]

> </soap:Envelope>

For individual one way transfers, this would relatively straightforward,
assuming that the relevant xml schema are well defined.  As a proof of
concept, I use separate one way transfers to sync my org files with
google's calendar.  Messy and ad hoc but it works; search the mailing list for
a description of how I did this -- I'm offline at the moment so cannot
search myself.  Alternatively, look at

[Worg]/org-tutorials/org-google-sync.html 

for a tutorial written by Arun Persaud that might give you an idea of
how this can be done.

HTH,
eric

-- 
: Eric S Fraga (GnuPG: 0xC89193D8FFFCF67D) in Emacs 24.0.90.1
: using Org-mode version 7.7 (release_7.7.452.g767f5)

How to debug "Specified time is not representable"

[Karl Voit]

Hi!

When I get «Specified time is not representable» while creating the
Agenda view, I want to get more information *where* the problem is.

I found [1] and following and so I got it that there is no way of
managing timestamps before 1970 :-(

Is there a way to get to the problematic time stamp?

  1. http://lists.gnu.org/archive/html/emacs-orgmode/2011-05/msg00729.html
-- 
Karl Voit

Re: How to debug "Specified time is not representable"

[Jambunathan K]

Karl Voit <devnull@Karl-Voit.at> writes:

> Hi!
>
> When I get «Specified time is not representable» while creating the
> Agenda view, I want to get more information *where* the problem is.
>
> I found [1] and following and so I got it that there is no way of
> managing timestamps before 1970 :-(
>
> Is there a way to get to the problematic time stamp?
>
>   1. http://lists.gnu.org/archive/html/emacs-orgmode/2011-05/msg00729.html

Hope you have looked at C-h v org-read-date-force-compatible-dates 

A simple M-x grep-find on .org files for the year should work.

You may also try

M-x debug-on-entry RET ding RET

Look at the backtrace and see whether you can get some clues.

Works best if your orgmode is not compiled

-- 

Re: How to debug "Specified time is not representable"

[Karl Voit]

Hi!

* Jambunathan K <kjambunathan@gmail.com> wrote:
> Karl Voit <devnull@Karl-Voit.at> writes:
>
>> When I get «Specified time is not representable» while creating the
>> Agenda view, I want to get more information *where* the problem is.
>>
>> I found [1] and following and so I got it that there is no way of
>> managing timestamps before 1970 :-(
>>
>> Is there a way to get to the problematic time stamp?
>>   1. http://lists.gnu.org/archive/html/emacs-orgmode/2011-05/msg00729.html
>
> Hope you have looked at C-h v org-read-date-force-compatible-dates 

Meanwhile: yes.

I do not want to use timestamps outside of UNIX epoch. I just wanted
to *locate* the culprit.

> A simple M-x grep-find on .org files for the year should work.

I do not know the year.

> You may also try
> M-x debug-on-entry RET ding RET
> Look at the backtrace and see whether you can get some clues.

Sorry, no clue.

> Works best if your orgmode is not compiled

When I have compiled Orgmode, should I delete all *.elc files?

-- 
Karl Voit

Re: How to debug "Specified time is not representable"

[Jambunathan K]

>> A simple M-x grep-find on .org files for the year should work.
>
> I do not know the year.

That is what regexps are for. 

If you visit the org file and do

M-x occur RET [<[]19[0-7][0-9] RET

you are likely to find all entries in the 1900 and 1979 range.

It is not my intention to provide you with a outright solution but only
to give sufficient hints so that you make progress.

>> You may also try
>> M-x debug-on-entry RET ding RET
>> Look at the backtrace and see whether you can get some clues.
>
> Sorry, no clue.
>
>> Works best if your orgmode is not compiled
>
> When I have compiled Orgmode, should I delete all *.elc files?

May be the debug-on-entry is the wrong strategy for the problem at
hand. You may disregard my suggestion.

-- 

Re: How to debug "Specified time is not representable"

[Karl Voit]

* Jambunathan K <kjambunathan@gmail.com> wrote:
>
>>> A simple M-x grep-find on .org files for the year should work.
>>
>> I do not know the year.
>
> That is what regexps are for. 
[...]
> It is not my intention to provide you with a outright solution but only
> to give sufficient hints so that you make progress.

:-)

Thanks. I know RegEx perfectly good. Irritating entries could be
located by...

    find . -name "*\.org*"|xargs egrep '<19[0-6]|<20[3-9]'

...since I am using *lots* of Orgmode files.

My point was more or less usability-related: what use is the message
«Specified time is not representable» when the user (anybody, not
just RegEx experts like us) gets no clue, where the problem is?

So I was curious, if there *is* some method I do not know (yet).

If the answer is «no, there is no way of telling you the actual time
stamp that causes the message», my question is answered. Bad for
Emacs/Orgmode usability but fine with me so far.

So is this really the case?

>>> You may also try
>>> M-x debug-on-entry RET ding RET
>>> Look at the backtrace and see whether you can get some clues.
>>
>> Sorry, no clue.
>>
>>> Works best if your orgmode is not compiled
>>
>> When I have compiled Orgmode, should I delete all *.elc files?
>
> May be the debug-on-entry is the wrong strategy for the problem at
> hand. You may disregard my suggestion.

-- 
Karl Voit

Re: How to debug "Specified time is not representable"

[Carsten Dominik]

On Oct 4, 2011, at 1:42 PM, Karl Voit wrote:

> * Jambunathan K <kjambunathan@gmail.com> wrote:
>> 
>>>> A simple M-x grep-find on .org files for the year should work.
>>> 
>>> I do not know the year.
>> 
>> That is what regexps are for. 
> [...]
>> It is not my intention to provide you with a outright solution but only
>> to give sufficient hints so that you make progress.
> 
> :-)
> 
> Thanks. I know RegEx perfectly good. Irritating entries could be
> located by...
> 
>    find . -name "*\.org*"|xargs egrep '<19[0-6]|<20[3-9]'
> 
> ...since I am using *lots* of Orgmode files.
> 
> My point was more or less usability-related: what use is the message
> «Specified time is not representable» when the user (anybody, not
> just RegEx experts like us) gets no clue, where the problem is?
> 
> So I was curious, if there *is* some method I do not know (yet).
> 
> If the answer is «no, there is no way of telling you the actual time
> stamp that causes the message», my question is answered. Bad for
> Emacs/Orgmode usability

Fair enough.  If you pull, you should now get decent error messages.

- Carsten


> but fine with me so far.
> 
> So is this really the case?
> 
>>>> You may also try
>>>> M-x debug-on-entry RET ding RET
>>>> Look at the backtrace and see whether you can get some clues.
>>> 
>>> Sorry, no clue.
>>> 
>>>> Works best if your orgmode is not compiled
>>> 
>>> When I have compiled Orgmode, should I delete all *.elc files?
>> 
>> May be the debug-on-entry is the wrong strategy for the problem at
>> hand. You may disregard my suggestion.
> 
> -- 
> Karl Voit
> 
> 

- Carsten

Re: How to debug "Specified time is not representable"

[Carsten Dominik]

On Oct 3, 2011, at 1:50 PM, Karl Voit wrote:

> Hi!
> 
> * Jambunathan K <kjambunathan@gmail.com> wrote:
>> Karl Voit <devnull@Karl-Voit.at> writes:
>> 
>>> When I get «Specified time is not representable» while creating the
>>> Agenda view, I want to get more information *where* the problem is.
>>> 
>>> I found [1] and following and so I got it that there is no way of
>>> managing timestamps before 1970 :-(
>>> 
>>> Is there a way to get to the problematic time stamp?
>>>  1. http://lists.gnu.org/archive/html/emacs-orgmode/2011-05/msg00729.html
>> 
>> Hope you have looked at C-h v org-read-date-force-compatible-dates 
> 
> Meanwhile: yes.
> 
> I do not want to use timestamps outside of UNIX epoch. I just wanted
> to *locate* the culprit.
> 
>> A simple M-x grep-find on .org files for the year should work.
> 
> I do not know the year.
> 
>> You may also try
>> M-x debug-on-entry RET ding RET
>> Look at the backtrace and see whether you can get some clues.
> 
> Sorry, no clue.


I have no general method.  However, you can do this:

1. On the "Options" menu, switch "Enter debug on error" on.
2. Run your agenda to hit the error.  A buffer with the backtrace
   pops open.
3. Press 

      e (current-buffer) RET

  to learn about the buffer where this happens

4. Press
   
      e pos RET

  to learn about the position of the offending time stamp.      

5. Press `q' to leave the backtrace buffer, then switch to the buffer
   you learned about and find the position you just learned about.
   THis should be the bad timestamp.

HTH

- Carsten


> 
>> Works best if your orgmode is not compiled
> 
> When I have compiled Orgmode, should I delete all *.elc files?
> 
> -- 
> Karl Voit
> 
> 

- Carsten

Re: How to debug "Specified time is not representable"

[Karl Voit]

Hi Carsten!

* Carsten Dominik <carsten.dominik@gmail.com> wrote:
>
> I have no general method.  However, you can do this:
>
> 1. On the "Options" menu, switch "Enter debug on error" on.
> 2. Run your agenda to hit the error.  A buffer with the backtrace
>    pops open.
> 3. Press 
>
>       e (current-buffer) RET
>
>   to learn about the buffer where this happens
>
> 4. Press
>    
>       e pos RET
>
>   to learn about the position of the offending time stamp.      
>
> 5. Press `q' to leave the backtrace buffer, then switch to the buffer
>    you learned about and find the position you just learned about.
>    THis should be the bad timestamp.

*This* is the method I was searching for. *Thank* you very much. Now
I can see the actual time stamp that causes any problem.

-- 
Karl Voit

Referencing elemts of a table

[Karl Voit]

Hi!

I want to define "orgtblB line 2 (row 1 to 6)" should contain the
values "from orgtblA, last line, row 2 to 7".

I tried

    #+TBLFM: @2=remote(orgtblA,@2$2..@2$7)

and 

    #+TBLFM: @2$1..@2$6=remote(orgtblA,@2$2..@2$7)

but that results in a list of 6x 5 values in each field in row 2.

Here is an example of my org-tbl-reference-problem:
http://paste.grml.org/462/

Thanks for your help!

-- 
Karl Voit

Re: Referencing elemts of a table

[Michael Brand]

Hi Karl

You need additionally $# from "Field coordinates in formulas" described here:
http://orgmode.org/manual/References.html#References
and Calc vector subscript:
#+TBLFM: @2 = subscr(remote(orgtblA, @2$2..@2$7), $#)

Michael

On Tue, Jun 7, 2011 at 17:55, Karl Voit <devnull@karl-voit.at> wrote:
> I want to define "orgtblB line 2 (row 1 to 6)" should contain the
> values "from orgtblA, last line, row 2 to 7".
>
> I tried
>
>    #+TBLFM: @2=remote(orgtblA,@2$2..@2$7)
>
> and
>
>    #+TBLFM: @2$1..@2$6=remote(orgtblA,@2$2..@2$7)
>
> but that results in a list of 6x 5 values in each field in row 2.
>
> Here is an example of my org-tbl-reference-problem:
> http://paste.grml.org/462/

Re: Referencing elemts of a table

[kinouchou]

[-- Attachment #1: Type: text/plain, Size: 998 bytes --]

Hello,

Do you know if it's possible for a table in another file? Because in the
manuel i read it's possible but I don't have result.

(sorry for my english)
Christelle


On Tue, Jun 7, 2011 at 7:22 PM, Michael Brand <michael.ch.brand@gmail.com>wrote:

> Hi Karl
>
> You need additionally $# from "Field coordinates in formulas" described
> here:
> http://orgmode.org/manual/References.html#References
> and Calc vector subscript:
> #+TBLFM: @2 = subscr(remote(orgtblA, @2$2..@2$7), $#)
>
> Michael
>
> On Tue, Jun 7, 2011 at 17:55, Karl Voit <devnull@karl-voit.at> wrote:
> > I want to define "orgtblB line 2 (row 1 to 6)" should contain the
> > values "from orgtblA, last line, row 2 to 7".
> >
> > I tried
> >
> >    #+TBLFM: @2=remote(orgtblA,@2$2..@2$7)
> >
> > and
> >
> >    #+TBLFM: @2$1..@2$6=remote(orgtblA,@2$2..@2$7)
> >
> > but that results in a list of 6x 5 values in each field in row 2.
> >
> > Here is an example of my org-tbl-reference-problem:
> > http://paste.grml.org/462/
>
>

[-- Attachment #2: Type: text/html, Size: 1715 bytes --]

Re: Referencing elemts of a table

[Jambunathan K]

> Do you know if it's possible for a table in another file? Because in
> the manuel i read it's possible but I don't have result.

M-x orgtbl-mode?


-- 

Re: Referencing elemts of a table

[Nick Dokos]

kinouchou <kinouchou@gmail.com> wrote:

> Do you know if it's possible for a table in another file? Because in the
> manuel i read it's possible but I don't have result.
> 
> (sorry for my english)
> Christelle
> 

I assume you mean the "remote references" in section 3.5.1 in the manual:

,----
| You may also reference constants, fields and ranges from a different
| table, either in the current file or even in a different file.  The
| syntax is
| 
|      remote(NAME-OR-ID,REF)
| 
| where NAME can be the name of a table in the current file as set by a
| `#+TBLNAME: NAME' line before the table.  It can also be the ID of an
| entry, even in a different file, and the reference then refers to the
| first table in that entry.  REF is an absolute field or range reference
| as described above for example `@3$3' or `$somename', valid in the
| referenced table.
`----

Here's a simple example - we have a table in one file, foo.org,
which refers to a cell in table with ID 17:
--8<---------------cut here---------------start------------->8---

* foo

#+TBLNAME: foo
|  a |   b |
|----+-----|
|  1 |   4 |
|  2 |   7 |
|  3 |  12 |
|  4 |  19 |
|  5 |  28 |
|  6 |  39 |
|  7 |  52 |
|  8 |  67 |
|  9 |  84 |
| 10 | 103 |

#+TBLFM: $2 = $1*$1 + remote(17, @4$1)
--8<---------------cut here---------------end--------------->8---

Here is the file that contains the table with ID 17 - the ID is
actually that of an entry and as the snippet of the manual says,
the remote reference refers to the first table under that entry.
Here we only have one table under the entry, so there is no
ambiguity:

--8<---------------cut here---------------start------------->8---


* bar
  :PROPERTIES:
  :ID:       17
  :END:

#+TBLNAME: bar
| a | b |
|---+---|
| 1 |   |
| 2 |   |
| 3 |   |
| 4 |   |
| 5 |   |
| 6 |   |
--8<---------------cut here---------------end--------------->8---

In practice, you would probably generate the ID with org-id-uuid
or similar to ensure that the ID is unique.

If you change the remote reference to e.g remote(17, @3$1) and
recalculate, you can see the values in the foo table changing
appropriately.

Does that answer your question?

Nick


> 
> On Tue, Jun 7, 2011 at 7:22 PM, Michael Brand <michael.ch.brand@gmail.com>wrote:
> 
> > Hi Karl
> >
> > You need additionally $# from "Field coordinates in formulas" described
> > here:
> > http://orgmode.org/manual/References.html#References
> > and Calc vector subscript:
> > #+TBLFM: @2 = subscr(remote(orgtblA, @2$2..@2$7), $#)
> >
> > Michael
> >
> > On Tue, Jun 7, 2011 at 17:55, Karl Voit <devnull@karl-voit.at> wrote:
> > > I want to define "orgtblB line 2 (row 1 to 6)" should contain the
> > > values "from orgtblA, last line, row 2 to 7".
> > >
> > > I tried
> > >
> > >    #+TBLFM: @2=remote(orgtblA,@2$2..@2$7)
> > >
> > > and
> > >
> > >    #+TBLFM: @2$1..@2$6=remote(orgtblA,@2$2..@2$7)
> > >
> > > but that results in a list of 6x 5 values in each field in row 2.
> > >
> > > Here is an example of my org-tbl-reference-problem:
> > > http://paste.grml.org/462/
> >
> >
> 

Re: Referencing elemts of a table

[Karl Voit]

* Michael Brand <michael.ch.brand@gmail.com> wrote:
> Hi Karl
>
> You need additionally $# from "Field coordinates in formulas" described here:
> http://orgmode.org/manual/References.html#References
> and Calc vector subscript:
> #+TBLFM: @2 = subscr(remote(orgtblA, @2$2..@2$7), $#)

This was the thing I was missing! Thank you!

But on the page of the URL mentioned above there is nothing
related to »subscr« at all. Is there a more verbose reference I do
not know yet?

-- 
Karl Voit

Re: Referencing elemts of a table

[Michael Brand]

Hi Karl

On Wed, Jun 8, 2011 at 14:49, Karl Voit <devnull@karl-voit.at> wrote:
> * Michael Brand <michael.ch.brand@gmail.com> wrote:
>> You need additionally $# from "Field coordinates in formulas" described here:
>> http://orgmode.org/manual/References.html#References
>> and Calc vector subscript:
>> #+TBLFM: @2 = subscr(remote(orgtblA, @2$2..@2$7), $#)
>
> This was the thing I was missing! Thank you!
>
> But on the page of the URL mentioned above there is nothing
> related to »subscr« at all. Is there a more verbose reference I do
> not know yet?

Org table spreadsheet gives access to Emacs Calc that has its own
manual and subscr is in chapter "Extracting Vector Elements":
(info "(Calc)Extracting Elements")

Michael

Re: Referencing elemts of a table

[Nick Dokos]

Karl Voit <devnull@Karl-Voit.at> wrote:

> * Michael Brand <michael.ch.brand@gmail.com> wrote:
> > Hi Karl
> >
> > You need additionally $# from "Field coordinates in formulas" described here:
> > http://orgmode.org/manual/References.html#References
> > and Calc vector subscript:
> > #+TBLFM: @2 = subscr(remote(orgtblA, @2$2..@2$7), $#)
> 
> This was the thing I was missing! Thank you!
> 
> But on the page of the URL mentioned above there is nothing
> related to »subscr« at all. Is there a more verbose reference I do
> not know yet?

Yes, it's documented in the Calc manual, section 11.3, "Extracting
Vector Elements". I'd encourage you to write up a short tutorial
on your use case: it's interesting, non-trivial and perhaps a nice
introduction to using Calc functions in Org tables. Worg would be
a better place for it!

Nick

org-time-stamp loses repeater interval

[Karl Voit]

Hi!

Maybe there is a (to me at least hidden) feature behind the behavior
that org-time-stamp (C-c .) deletes any repeater information when
used to update a date stamp.

Or is this some kind of bug or at least unexpected behavior?

-- 
Karl Voit

Re: org-time-stamp loses repeater interval

[Bernt Hansen]

Karl Voit <devnull@Karl-Voit.at> writes:

> Hi!
>
> Maybe there is a (to me at least hidden) feature behind the behavior
> that org-time-stamp (C-c .) deletes any repeater information when
> used to update a date stamp.
>
> Or is this some kind of bug or at least unexpected behavior?

There was a report of that recently for malformed timestamps.  Do you
have an example of one that loses the repeater when you reschedule it?

-- 
Bernt

Re: org-time-stamp loses repeater interval

[Nick Dokos]

Karl Voit <devnull@Karl-Voit.at> wrote:


> Maybe there is a (to me at least hidden) feature behind the behavior
> that org-time-stamp (C-c .) deletes any repeater information when
> used to update a date stamp.
> 

AFAIK, org-time-stamp creates a brand-new time stamp: it does not update
an existing one. You can update the time stamp by just using ordinary
Emacs command ("it's just text"), or you can put the cursor on various
parts of the time stamp and do S-up, S-down (bound to org-shiftup and
org-shiftdown resp.) to change that part.

Nick

Re: org-time-stamp loses repeater interval

[Michael Brand]

On Fri, Jun 24, 2011 at 18:01, Nick Dokos <nicholas.dokos@hp.com> wrote:
> Karl Voit <devnull@Karl-Voit.at> wrote:
>> Maybe there is a (to me at least hidden) feature behind the behavior
>> that org-time-stamp (C-c .) deletes any repeater information when
>> used to update a date stamp.

Same here with "DEADLINE: <2011-06-25 Sat +1w>"

> AFAIK, org-time-stamp creates a brand-new time stamp: it does not update
> an existing one. You can update the time stamp by just using ordinary
> Emacs command ("it's just text"), or you can put the cursor on various
> parts of the time stamp and do S-up, S-down (bound to org-shiftup and
> org-shiftdown resp.) to change that part.

If it is a DEADLINE or SCHEDULED you can also use "C-c C-d ." or "C-c
C-s ." as a workaround to preserve the repeater. Therefore I consider
loosing the repeater with just "C-c ." on any active timestamp, no
matter if a DEADLINE, SCHEDULED or not, a bug.

Michael

Re: org-time-stamp loses repeater interval

[Nick Dokos]

Michael Brand <michael.ch.brand@gmail.com> wrote:

> On Fri, Jun 24, 2011 at 18:01, Nick Dokos <nicholas.dokos@hp.com> wrote:
> > Karl Voit <devnull@Karl-Voit.at> wrote:
> >> Maybe there is a (to me at least hidden) feature behind the behavior
> >> that org-time-stamp (C-c .) deletes any repeater information when
> >> used to update a date stamp.
> 
> Same here with "DEADLINE: <2011-06-25 Sat +1w>"
> 
> > AFAIK, org-time-stamp creates a brand-new time stamp: it does not update
> > an existing one. You can update the time stamp by just using ordinary
> > Emacs command ("it's just text"), or you can put the cursor on various
> > parts of the time stamp and do S-up, S-down (bound to org-shiftup and
> > org-shiftdown resp.) to change that part.
> 
> If it is a DEADLINE or SCHEDULED you can also use "C-c C-d ." or "C-c
> C-s ." as a workaround to preserve the repeater. Therefore I consider
> loosing the repeater with just "C-c ." on any active timestamp, no
> matter if a DEADLINE, SCHEDULED or not, a bug.
> 

I stand corrected: it does sound like a bug.

Nick

Re: Bug: org-time-stamp loses repeater interval

[Karl Voit]

* Michael Brand <michael.ch.brand@gmail.com> wrote:
>
> If it is a DEADLINE or SCHEDULED you can also use "C-c C-d ." or "C-c
> C-s ." as a workaround to preserve the repeater. Therefore I consider
> loosing the repeater with just "C-c ." on any active timestamp, no
> matter if a DEADLINE, SCHEDULED or not, a bug.

Same to me. Unfortunately I am using timestamps not only in the
context of DEADLINE or SCHEDULED but also for events that should
simply show up on the agenda (without deadline or scheduled
timestamp at all).

To me this *is* a bug since my expected behavior of org-time-stamp
to modify an existing timestamp is different.

So, how reaches this bug report a person who is able to fix it?


PS: Sorry, I still don't know ELISP yet and therefore can't fix it
by myself.

-- 
Karl Voit

Re: Bug: org-time-stamp loses repeater interval

[Bastien]

Karl Voit <devnull@Karl-Voit.at> writes:

> Same to me. Unfortunately I am using timestamps not only in the
> context of DEADLINE or SCHEDULED but also for events that should
> simply show up on the agenda (without deadline or scheduled
> timestamp at all).
>
> To me this *is* a bug since my expected behavior of org-time-stamp
> to modify an existing timestamp is different.

Yes, this is a bug.

> So, how reaches this bug report a person who is able to fix it?

You've just done so :)  

It is now fixed in the latest git version.  Please test and report, 
if possible.

> PS: Sorry, I still don't know ELISP yet and therefore can't fix it
> by myself.

No problem!

-- 
 Bastien

Re: Bug: org-time-stamp loses repeater interval

[Karl Voit]

* Bastien <bzg@altern.org> wrote:
> Karl Voit <devnull@Karl-Voit.at> writes:
>
>> Same to me. Unfortunately I am using timestamps not only in the
>> context of DEADLINE or SCHEDULED but also for events that should
>> simply show up on the agenda (without deadline or scheduled
>> timestamp at all).
>>
>> To me this *is* a bug since my expected behavior of org-time-stamp
>> to modify an existing timestamp is different.
>
> It is now fixed in the latest git version.  Please test and report, 
> if possible.

I checked out d7c693c1ffe2ebf38604dd30fdd2f9d51a1c44d0 which should
fix the problem and tested with «Emacs 23.2.1 (i686-pc-linux-gnu,
GTK+ Version 2.24.4) of 2011-04-04 on rothera, modified by Debian»
and «Org-mode version 7.5 (release_7.5.462.gd7c69)»

created timestamp with «C-c .»
<2011-06-28 Tue>

modified with «C-c.» to wednesday:
<2011-06-29 Wed>

manually added repeater:
<2011-06-29 Wed +1w>

«C-c .» + click on Jul 5th:
<2011-07-05 Tue>

-> repeater is lost :-(

created another one with warning period:
<2011-06-28 Tue -1d>

«C-c .» + click on Jul 5th:
<2011-07-05 Tue>

-> warning period is lost :-(

created another one with repeater and warning period:
<2011-06-28 Tue +1w -1d>

«C-c .» + click on Jul 5th:
<2011-07-05 Tue>

-> repeater and warning period is lost :-(

-- 
Karl Voit

Re: Bug: org-time-stamp loses repeater interval

[Bastien]

Karl Voit <devnull@Karl-Voit.at> writes:

> -> repeater and warning period is lost :-(

I cannot reproduce this. 

Did you take care of loading the freshly pulled Org version?

-- 
 Bastien

Re: Bug: org-time-stamp loses repeater interval

[Nick Dokos]

Bastien <bzg@altern.org> wrote:

> Karl Voit <devnull@Karl-Voit.at> writes:
> 
> > -> repeater and warning period is lost :-(
> 
> I cannot reproduce this. 
> 
> Did you take care of loading the freshly pulled Org version?
> 

I can't reproduce this either. Both repeaters and warning periods
are preserved with C-c . (and with C-c C-s and C-c C-d as well).

Nick

Re: Bug: org-time-stamp loses repeater interval

[Karl Voit]

* Bastien <bzg@altern.org> wrote:
> Karl Voit <devnull@Karl-Voit.at> writes:
>
>> -> repeater and warning period is lost :-(
>
> I cannot reproduce this. 
> Did you take care of loading the freshly pulled Org version?

«git pull» on the master and then I re-started my Emacs. I thought
this should be enough ...

Or is there any setting that can conflict and produce the behavior
of my Org-mode?

-- 
Karl Voit

Re: Bug: org-time-stamp loses repeater interval

[Nick Dokos]

Karl Voit <devnull@Karl-Voit.at> wrote:

> * Bastien <bzg@altern.org> wrote:
> > Karl Voit <devnull@Karl-Voit.at> writes:
> >
> >> -> repeater and warning period is lost :-(
> >
> > I cannot reproduce this. 
> > Did you take care of loading the freshly pulled Org version?
> 
> «git pull» on the master and then I re-started my Emacs. I thought
> this should be enough ...
> 
> Or is there any setting that can conflict and produce the behavior
> of my Org-mode?
> 

Do you use compiled (.elc) files? If so, you need to "make clean; make"
before restarting. You can also check with

M-x locate-library RET org RET

to see what would get loaded.

Nick

Re: Bug: org-time-stamp loses repeater interval

[Karl Voit]

* Nick Dokos <nicholas.dokos@hp.com> wrote:
>
> Do you use compiled (.elc) files? 

Not intentionally.

> If so, you need to "make clean; make" before restarting. You can
> also check with

This one did the trick. Thanks!

-- 
Karl Voit

Re: Bug: org-time-stamp loses repeater interval

[Bastien]

Karl Voit <devnull@Karl-Voit.at> writes:

> «git pull» on the master and then I re-started my Emacs. I thought
> this should be enough ...

If you byte-compiled the previous version of Org, Emacs will load 
this one, not the fresh one.

> Or is there any setting that can conflict and produce the behavior
> of my Org-mode?

Try M-x org-reload RET to explicitely reload Org. 

-- 
 Bastien

Re: Bug: org-time-stamp loses repeater interval

[Karl Voit]

After re-compiling my elc files, I have now the following behavior:

* Karl Voit <devnull@Karl-Voit.at> wrote:
>
> created timestamp with «C-c .»
> <2011-06-28 Tue>
>
> modified with «C-c.» to wednesday:
> <2011-06-29 Wed>
>
> manually added repeater:
> <2011-06-29 Wed +1w>
>
> «C-c .» + click on Jul 5th:

<2011-07-05 Tue +1w>

... which is the fix to the original bug. Thanks! :-)

> created another one with warning period:
> <2011-06-28 Tue -1d>
>
> «C-c .» + click on Jul 5th:
> <2011-07-05 Tue>
>
> -> warning period is lost :-(

The warning period still gets deleted though :-(

-- 
Karl Voit

Re: Bug: org-time-stamp loses repeater interval

[Bastien]

Karl Voit <devnull@Karl-Voit.at> writes:

> The warning period still gets deleted though :-(

You're right, should be fixed now.

Thanks!

-- 
 Bastien

Re: Bug: org-time-stamp loses repeater interval

[Karl Voit]

* Bastien <bzg@altern.org> wrote:
> Karl Voit <devnull@Karl-Voit.at> writes:
>
>> The warning period still gets deleted though :-(
>
> You're right, should be fixed now.

Sorry when I disagree for one case:

When I change each entry in my test data using «C-c .» and clicking
on 1st of July ...

,----[ test data ]
| <2011-06-28 Tue>
| <2011-06-28 Tue +1w>
| <2011-06-28 Tue -1d>
| <2011-06-28 Tue +1w -1d>
`----

... I end up having this:

,----[ result ]
| <2011-07-01 Fri>
| <2011-07-01 Fri +1w>
| <2011-07-01 Fri -1d>
| <2011-07-01 Fri -1d>
`----

In the last case, the repeater gets lost :-(

-- 
Karl Voit

Hallo, mein Name ist Web 2.0. Gib mir dein Adressbuch und lass mich Emails lesen

Re: Bug: org-time-stamp loses repeater interval

[Bastien]

Hello Karl,

Karl Voit <devnull@Karl-Voit.at> writes:

> Sorry when I disagree for one case:
>
> When I change each entry in my test data using «C-c .» and clicking
> on 1st of July ...
>
> ,----[ test data ]
> | <2011-06-28 Tue>
> | <2011-06-28 Tue +1w>
> | <2011-06-28 Tue -1d>
> | <2011-06-28 Tue +1w -1d>
> `----
>
> ... I end up having this:
>
> ,----[ result ]
> | <2011-07-01 Fri>
> | <2011-07-01 Fri +1w>
> | <2011-07-01 Fri -1d>
> | <2011-07-01 Fri -1d>
> `----
>
> In the last case, the repeater gets lost :-(

I cannot reproduce this.

Note that if you use the warning _before_ the repeater,
then it gets losts: 

<2011-06-28 Tue -1d +1w>

That because the warning should always be _after_ the repeater:

See section 8.3.2 "Repeated tasks" in the manual

  "the repeater should come first and the warning period last:
  `DEADLINE: <2005-10-01 Sat +1m -3d>'."

HTH,

-- 
 Bastien

Re: Bug: org-time-stamp loses repeater interval

[Nick Dokos]

Bastien <bzg@altern.org> wrote:

> Hello Karl,
> 
> Karl Voit <devnull@Karl-Voit.at> writes:
> 
> > Sorry when I disagree for one case:
> >
> > When I change each entry in my test data using «C-c .» and clicking
> > on 1st of July ...
> >
> > ,----[ test data ]
> > | <2011-06-28 Tue>
> > | <2011-06-28 Tue +1w>
> > | <2011-06-28 Tue -1d>
> > | <2011-06-28 Tue +1w -1d>
> > `----
> >
> > ... I end up having this:
> >
> > ,----[ result ]
> > | <2011-07-01 Fri>
> > | <2011-07-01 Fri +1w>
> > | <2011-07-01 Fri -1d>
> > | <2011-07-01 Fri -1d>
> > `----
> >
> > In the last case, the repeater gets lost :-(
> 
> I cannot reproduce this.
> 

I can - but after earlier today, I'm not sure I trust my eyes: anybody else?

Org-mode version 7.5 (baseline.387.gf9cc.dirty)

(the dirty part is because of some Makefile experiments, nothing to do with timestamps).

> Note that if you use the warning _before_ the repeater,
> then it gets losts: 
> 
> <2011-06-28 Tue -1d +1w>

That's not the problem - check the "before" state of the third of Karl's
examples.

> 
> That because the warning should always be _after_ the repeater:
> 
> See section 8.3.2 "Repeated tasks" in the manual
> 
>   "the repeater should come first and the warning period last:
>   `DEADLINE: <2005-10-01 Sat +1m -3d>'."
> 

Nick

Re: Bug: org-time-stamp loses repeater interval

[Bastien]

Okay, I've pushed another fix.

This let me stumble upon another case: the one with org-schedule and
org-deadline ignoring warning cookies -- these cases are also fixed.

Please confirm!

Thanks,

-- 
 Bastien

Re: Bug: org-time-stamp loses repeater interval

[Nick Dokos]

Bastien <bzg@altern.org> wrote:

> Okay, I've pushed another fix.
> 
> This let me stumble upon another case: the one with org-schedule and
> org-deadline ignoring warning cookies -- these cases are also fixed.
> 
> Please confirm!
> 

Confirmed. There is a peculiar corner case:

If I have a headline that's both scheduled and deadlined, like this:

--8<---------------cut here---------------start------------->8---
* scheduled 
  DEADLINE: <2011-07-04 Mon +2w -3d>
  SCHEDULED: <2011-07-06 Wed +1w -2d>
--8<---------------cut here---------------end--------------->8---

and I C-c C-s in the scheduled date, I get a second SCHEDULED: item
with the new date on the DEADLINE line. The original SCHEDULED: is
still on the next line, unchanged - like this:

--8<---------------cut here---------------start------------->8---
* scheduled 
  DEADLINE: <2011-07-04 Mon +2w -3d> SCHEDULED: <2011-07-03 Sun +1w -2d>
  SCHEDULED: <2011-07-06 Wed +1w -2d>
--8<---------------cut here---------------end--------------->8---

Nick

Org-mode version 7.5 (baseline.403.g13573)

Re: Bug: org-time-stamp loses repeater interval

[Sebastien Vauban]

Hi Nick,

Nick Dokos wrote:
> Bastien <bzg-whniv8GeeGkdnm+yROfE0A@public.gmane.org> wrote:
>> Okay, I've pushed another fix.
>> 
>> This let me stumble upon another case: the one with org-schedule and
>> org-deadline ignoring warning cookies -- these cases are also fixed.
>> 
>> Please confirm!
>
> Confirmed. There is a peculiar corner case:
>
> If I have a headline that's both scheduled and deadlined, like this:
>
> * scheduled 
>   DEADLINE: <2011-07-04 Mon +2w -3d>
>   SCHEDULED: <2011-07-06 Wed +1w -2d>
>
> and I C-c C-s in the scheduled date, I get a second SCHEDULED: item
> with the new date on the DEADLINE line. The original SCHEDULED: is
> still on the next line, unchanged - like this:
>
> * scheduled 
>   DEADLINE: <2011-07-04 Mon +2w -3d> SCHEDULED: <2011-07-03 Sun +1w -2d>
>   SCHEDULED: <2011-07-06 Wed +1w -2d>

See http://www.mail-archive.com/emacs-orgmode-mXXj517/zsQ@public.gmane.org/msg37987.html where I
report such a case with inactive timestamps and SCHEDULED dates.

See Bastien's answer in the same thread. In this case, SCHEDULED should come
first, before DEADLINE, for it to work.

Note -- I prefer that order (SCHEDULED, then DEADLINE) since dates are then
chronologically sorted (at least, I expect so, that SCHEDULED date < DEADLINE
date)...

Best regards,
  Seb

-- 
Sebastien Vauban

Re: Bug: org-time-stamp loses repeater interval

[Bastien]

Hi Sebastien,

"Sebastien Vauban" <wxhgmqzgwmuf@spammotel.com> writes:

> See http://www.mail-archive.com/emacs-orgmode@gnu.org/msg37987.html where I
> report such a case with inactive timestamps and SCHEDULED dates.
>
> See Bastien's answer in the same thread. In this case, SCHEDULED should come
> first, before DEADLINE, for it to work.

There are still strange things here, I want to address them before 7.6.

Stay tuned.

-- 
 Bastien

Re: Bug: org-time-stamp loses repeater interval

[Nick Dokos]

Sebastien Vauban <wxhgmqzgwmuf@spammotel.com> wrote:

> Hi Nick,
> 
> Nick Dokos wrote:
> > Bastien <bzg@altern.org> wrote:
> >> Okay, I've pushed another fix.
> >> 
> >> This let me stumble upon another case: the one with org-schedule and
> >> org-deadline ignoring warning cookies -- these cases are also fixed.
> >> 
> >> Please confirm!
> >
> > Confirmed. There is a peculiar corner case:
> >
> > If I have a headline that's both scheduled and deadlined, like this:
> >
> > * scheduled 
> >   DEADLINE: <2011-07-04 Mon +2w -3d>
> >   SCHEDULED: <2011-07-06 Wed +1w -2d>
> >
> > and I C-c C-s in the scheduled date, I get a second SCHEDULED: item
> > with the new date on the DEADLINE line. The original SCHEDULED: is
> > still on the next line, unchanged - like this:
> >
> > * scheduled 
> >   DEADLINE: <2011-07-04 Mon +2w -3d> SCHEDULED: <2011-07-03 Sun +1w -2d>
> >   SCHEDULED: <2011-07-06 Wed +1w -2d>
> 
> See http://www.mail-archive.com/emacs-orgmode@gnu.org/msg37987.html where I
> report such a case with inactive timestamps and SCHEDULED dates.
> 
> See Bastien's answer in the same thread. In this case, SCHEDULED should come
> first, before DEADLINE, for it to work.
> 
> Note -- I prefer that order (SCHEDULED, then DEADLINE) since dates are then
> chronologically sorted (at least, I expect so, that SCHEDULED date < DEADLINE
> date)...
> 

Seb,

thanks for pointing out the thread. I think Bastien is referring to the
order of {SCHEDULED or DEADLINE} against inactive, so it's not quite the
same - in particular, he mentions that he uses SCHEDULED and DEADLINE
together, but I cannot see a dictum about which one of those should come
first.

For my part, my needs are simple: I never have more than one timestamp
on an item, so I don't run into these corner cases usually (except when
I'm in twisted testing mode). And the existence of these corner cases
validates my approach[fn:1].

It would theoretically be better if there were no rules of course: you
could put any set of timestamps in any order you liked and org would do
the right thing in all cases. But I doubt very much that the effort is
worth it. If there are hard-and-fast rules, about ordering however, an
addition to the docs might be a good idea (unless it's already there -
I haven't checked).

The other thing that I *think* I ran into is that occasionally, with a
DEADLINE and SCHEDULED on the same line, changing one would change the
*order*. I did wonder whether org was chronologically ordering them, but
that was not the case. However, I cannot duplicate the reordering at all
now (I tried quite a few times), so I could have imagined it (after all,
it happened yesterday, which was a red-letter day of imaginary findings
for me).

Thanks,
Nick

Footnotes:

[fn:1] :-)

Re: Bug: org-time-stamp loses repeater interval

[Bastien]

Hi Nick,

Nick Dokos <nicholas.dokos@hp.com> writes:

> The other thing that I *think* I ran into is that occasionally, with a
> DEADLINE and SCHEDULED on the same line, changing one would change the
> *order*. I did wonder whether org was chronologically ordering them, but
> that was not the case. 

It is now -- most recent timestamp being put first, on the front of the
line.

Of course, I hope one day DEADLINE and SCHEDULED will just be ordinary 
properties...  so that we can use the org-set-property API instead of
going down all the hassle of positioning things "manually" (please don't
look at the code of `org-add-planning-info'!!)

Best,

-- 
 Bastien

Re: Bug: org-time-stamp loses repeater interval

[Karl Voit]

I checked out via «git pull», re-compiled Org-mode and tested again:
this time, all my test cases mentioned in the original posting
worked fine. Bug fixed so far.

* Nick Dokos <nicholas.dokos@hp.com> wrote:
>
> There is a peculiar corner case:
>
> If I have a headline that's both scheduled and deadlined, like this:
>
> --8<---------------cut here---------------start------------->8---
> * scheduled 
>   DEADLINE: <2011-07-04 Mon +2w -3d>
>   SCHEDULED: <2011-07-06 Wed +1w -2d>
> --8<---------------cut here---------------end--------------->8---
>
> and I C-c C-s in the scheduled date, I get a second SCHEDULED: item
> with the new date on the DEADLINE line. The original SCHEDULED: is
> still on the next line, unchanged - like this:
>
> --8<---------------cut here---------------start------------->8---
> * scheduled 
>   DEADLINE: <2011-07-04 Mon +2w -3d> SCHEDULED: <2011-07-03 Sun +1w -2d>
>   SCHEDULED: <2011-07-06 Wed +1w -2d>
> --8<---------------cut here---------------end--------------->8---

I can confirm this behavior at my side too.

But: When I create SCHEDULED (and DEADLINE) using the shortcut «C-c
C-s» SCHEDULED is always in the same like as DEADLINE. In this case,
I can not notice the behavior as above. The behavior described by
Nick happens only (here) if SCHEDULED is not in the same like as
DEADLINE.

Remark regarding the order of the two entries: the last thing that
gets added/modified is always the last one in this line. So if I
create SCHEDULED and DEADLINE, DEADLINE is the second (last) in the
line. When I modify SCHEDULED afterwards, it is SCHEDULED that gets
positioned at the end of the line. Same for modifications of
DEADLINE.

-- 
Karl Voit

Re: Bug: org-time-stamp loses repeater interval

[Bastien]

Hi Karl,

Karl Voit <devnull@Karl-Voit.at> writes:

> Remark regarding the order of the two entries: the last thing that
> gets added/modified is always the last one in this line. So if I
> create SCHEDULED and DEADLINE, DEADLINE is the second (last) in the
> line. When I modify SCHEDULED afterwards, it is SCHEDULED that gets
> positioned at the end of the line. Same for modifications of
> DEADLINE.

I've pushed fixes so that org-time-stamp, org-scheduled and org-deadline
put timestamps in a more stable way.  

Please check against the latest git repo and report any problem.

Thanks!

-- 
 Bastien

Re: Bug: org-time-stamp loses repeater interval

[Bastien]

I want to enforce a policy of having SCHEDULED, DEADLINED and CLOSED
always on the *same line* -- which is the default behavior when using 
the command.  

Would that break too many .org files?

-- 
 Bastien

Re: How to debug "Specified time is not representable"

[Nick Dokos]

Karl Voit <devnull@Karl-Voit.at> wrote:

> Hi!
> 
> When I get «Specified time is not representable» while creating the
> Agenda view, I want to get more information *where* the problem is.
> 
> I found [1] and following and so I got it that there is no way of
> managing timestamps before 1970 :-(
> 

Not quite true: Carsten introduced org-read-date-force-compatible-dates
which allows the user to take full responsibility for shooting
{him,her}self in the foot :-)

See 

    http://thread.gmane.org/gmane.emacs.orgmode/39206

You can try setting it to nil and see if your system dtrt, but you
should read the caveats in the doc string of the variable.

Nick

> Is there a way to get to the problematic time stamp?
> 
>   1. http://lists.gnu.org/archive/html/emacs-orgmode/2011-05/msg00729.html
> -- 
> Karl Voit
> 
> 

Re: Recurring events with exceptions

[Nick Dokos]

Karl Voit <devnull@Karl-Voit.at> wrote:

> Hi!
> 
> I am into a process to write a convert tool from my old calendar
> software[1] to Org-mode.
> 
> Now I do have to define something like »this event is recurring each
> week on Wednesday except 2011-10-26 and 2011-11-30«.
> 
> I already know that complex things have to be done using sexp
> entries[2] but this does not seem to be possible with sexp either.
> 

Why not? All you have to do is explicitly exclude the two
dates. Something like this (lightly tested) should work.  (NB: Wednesday
= 3 and calendar/diary wants dates as 3-element lists in the form (month
day year)):


%%(let ((dayname (calendar-day-of-week date)))
	(and (= dayname 3)
	     (not (calendar-date-equal date '(10 26 2011)))
	     (not (calendar-date-equal date '(11 30 2011)))))

I added this to .diary, did M-x calendar RET and marked the entries
with m: seems to work fine.

Nick

> Before I do have to develop a method that generates multiple
> distinct events for each recurring definition: is there another way
> to achieve this?
> 
> Thanks!
> 
>   1. jPilot/DateBK6/PalmOS
>   2. http://www.gnu.org/software/emacs/manual/html_node/emacs/Sexp-Diary-Entries.html
> -- 
> Karl Voit
> 
> 

Re: Recurring events with exceptions

[Karl Voit]

* Nick Dokos <nicholas.dokos@hp.com> wrote:
>
> All you have to do is explicitly exclude the two
> dates. Something like this (lightly tested) should work.  (NB: Wednesday
> = 3 and calendar/diary wants dates as 3-element lists in the form (month
> day year)):
>
> %%(let ((dayname (calendar-day-of-week date)))
> 	(and (= dayname 3)
> 	     (not (calendar-date-equal date '(10 26 2011)))
> 	     (not (calendar-date-equal date '(11 30 2011)))))
>
> I added this to .diary, did M-x calendar RET and marked the entries
> with m: seems to work fine.

Oh, cool. Thanks for the example!

-- 
Karl Voit

Re: [bug] org-agenda-write does not handle date stamps without day of week

[Nick Dokos]

Karl Voit <devnull@Karl-Voit.at> wrote:

> Since no followup reaction occured: is this a but that requires much
> work? Or did my email just got lost in overfull inboxes? :-)
> 

For me, it was a "no time to work on org - stash it"...

> * Karl Voit <devnull@Karl-Voit.at> wrote:
> >
> > I verified the behavior below with the most current git version of
> > Org-mode.
> >
> > Whenever I call my function to export the current agenda, it
> > converts that do have a beginning and end time to events
> > that span the whole day (without begin and end times at all).
> >
> > ,----[ my LISP command to export ]
> >| (defun vk-export-agenda()
> >|   "Exports monthly Org-mode agenda to agenda.ics file"
> >|   (interactive)
> >|   (org-agenda-list nil nil 60)
> >|   (org-agenda-write "~/org/agenda.ics")
> >| )
> > `----
> >
> > I traced down the problem to this:
> >
> > * <2012-03-05 Mon 08:00-09:00> works fine
> > * <2012-03-05 Mon 8:00-9:00> works fine too
> >
> > * <2012-03-05 08:00-09:00> Wrong: ends up as full day event
> >
> > Unfortunately, I do have certain mechanisms that generate parts of
> > my Org-mode files where I skip the day of the week. Org-mode in
> > general is able to handle date stamps without day of week pretty
> > well. But whenever I export it to ics (in order to send it to
> > Google), I end up with a messed up online agenda.
> >
> > I guess that org-agenda-write is the one to blame. It would be very
> > handy to me if somebody could fix that issue.

org-agenda-write calls org-export-icalendar which calls org-print-icalendar-entries
which loops over all the entries and parses them, decomposing them into timestamps.
Each timestamp is then passed to org-parse-time-string. It's this one that cannot
handle non-standard formats: it uses a regexp and assumes that all the matched parts
are going to be in fixed places:

,----
| (defun org-parse-time-string (s &optional nodefault)
|   "Parse the standard Org-mode time string.
| This should be a lot faster than the normal `parse-time-string'.
| If time is not given, defaults to 0:00.  However, with optional NODEFAULT,
| hour and minute fields will be nil if not given."
|   (if (string-match org-ts-regexp0 s)
|       (list 0
| 	    (if (or (match-beginning 8) (not nodefault))
| 		(string-to-number (or (match-string 8 s) "0")))
| 	    (if (or (match-beginning 7) (not nodefault))
| 		(string-to-number (or (match-string 7 s) "0")))
| 	    (string-to-number (match-string 4 s))
| 	    (string-to-number (match-string 3 s))
| 	    (string-to-number (match-string 2 s))
| 	    nil nil nil)
|     (error "Not a standard Org-mode time string: %s" s)))
`----

You can evaluate the following in your *scratch* buffer to verify:

--8<---------------cut here---------------start------------->8---
(org-parse-time-string "<2012-03-05 Mon 08:00>")
(0 0 8 5 3 2012 nil nil nil)

(org-parse-time-string "<2012-03-05 08:00>")
(0 0 0 5 3 2012 nil nil nil)
--8<---------------cut here---------------end--------------->8---

As you can see, the time has disappeared.

As to how to fix it, there are several possibilities:

1. fix your scripts that produce time stamps to include day-of-week.
2. change the callers of org-parse-time-string to make sure that DOW is included.
3. change just one caller: org-print-icalendar-entries to make sure that DOW is included.
4. change org-parse-time-string to handle a missing DOW.

There are roughly three dozen callers, so 2. is possible but a pain.
3. is simple but ugly as sin, 4. is the best way to handle it within org.

I vote for 1. where *you* have to do all the work ;-)

Nick

Re: [bug] org-agenda-write does not handle date stamps without day of week

[Karl Voit]

* Nick Dokos <nicholas.dokos@hp.com> wrote:
> Karl Voit <devnull@Karl-Voit.at> wrote:
>
> For me, it was a "no time to work on org - stash it"...

OK. I just wanted to make sure that it *is* on someone's todo list
:-)

>> * Karl Voit <devnull@Karl-Voit.at> wrote:
>> >
>> > * <2012-03-05 08:00-09:00> Wrong: ends up as full day event
>
> org-agenda-write calls org-export-icalendar which calls org-print-icalendar-entries
> which loops over all the entries and parses them, decomposing them into timestamps.
> Each timestamp is then passed to org-parse-time-string. It's this one that cannot
> handle non-standard formats: it uses a regexp and assumes that all the matched parts
> are going to be in fixed places:
>
> As to how to fix it, there are several possibilities:
>
> 1. fix your scripts that produce time stamps to include day-of-week.

Sorry, deriving DOW from an arbitrary timestamps from arbitrary data
sources is either pretty time consuming (calendar calculations) or
simply hard to calculate.

Outside Org-mode, DOW is seldom part of time-stamp data :-(

> 2. change the callers of org-parse-time-string to make sure that DOW is included.
> There are roughly three dozen callers, so 2. is possible but a pain.

Ack.

> 3. change just one caller: org-print-icalendar-entries to make sure that DOW is included.
> 3. is simple but ugly as sin, 

Ouch, ack :-)

> 4. change org-parse-time-string to handle a missing DOW.
> 4. is the best way to handle it within org.

Full ack.

> I vote for 1. where *you* have to do all the work ;-)

YMMD :-)

If my brain would be compatible to ELISP, I'd send a patch.
Promised.

But I'll take my chance and wait for someone else (you?)
implementing 4. to resolve this issue for everybody. I really
appreciate every second you guys invest in maintaining Org-mode!

-- 
Karl Voit

Re: [bug] org-agenda-write does not handle date stamps without day of week

[Nick Dokos]

[-- Attachment #1: Type: text/plain, Size: 2854 bytes --]

Karl Voit <devnull@Karl-Voit.at> wrote:

> * Nick Dokos <nicholas.dokos@hp.com> wrote:
> > Karl Voit <devnull@Karl-Voit.at> wrote:
> >
> > For me, it was a "no time to work on org - stash it"...
> 
> OK. I just wanted to make sure that it *is* on someone's todo list
> :-)
> 
> >> * Karl Voit <devnull@Karl-Voit.at> wrote:
> >> >
> >> > * <2012-03-05 08:00-09:00> Wrong: ends up as full day event
> >
> > org-agenda-write calls org-export-icalendar which calls org-print-icalendar-entries
> > which loops over all the entries and parses them, decomposing them into timestamps.
> > Each timestamp is then passed to org-parse-time-string. It's this one that cannot
> > handle non-standard formats: it uses a regexp and assumes that all the matched parts
> > are going to be in fixed places:
> >
> > As to how to fix it, there are several possibilities:
> >
> > 1. fix your scripts that produce time stamps to include day-of-week.
> 
> Sorry, deriving DOW from an arbitrary timestamps from arbitrary data
> sources is either pretty time consuming (calendar calculations) or
> simply hard to calculate.
> 
> Outside Org-mode, DOW is seldom part of time-stamp data :-(
> 
> > 2. change the callers of org-parse-time-string to make sure that DOW is included.
> > There are roughly three dozen callers, so 2. is possible but a pain.
> 
> Ack.
> 
> > 3. change just one caller: org-print-icalendar-entries to make sure that DOW is included.
> > 3. is simple but ugly as sin, 
> 
> Ouch, ack :-)
> 
> > 4. change org-parse-time-string to handle a missing DOW.
> > 4. is the best way to handle it within org.
> 
> Full ack.
> 
> > I vote for 1. where *you* have to do all the work ;-)
> 
> YMMD :-)
> 
> If my brain would be compatible to ELISP, I'd send a patch.
> Promised.
> 
> But I'll take my chance and wait for someone else (you?)
> implementing 4. to resolve this issue for everybody. I really
> appreciate every second you guys invest in maintaining Org-mode!
> 

I don't know about you, but whenever I engage in hand-to-hand combat
with a complicated regexp, I come out bruised, muddied and a lot worse
for wear. In any case, I'm attaching an org file with my investigations.
It contains a description and a code block for testing.

I hope that the attachment will come through unscathed: it contains
regexps, and munging a regexp that looks like hen scratchings in the
first place through uncooperative mailers is not something to be
relished.

BTW, I'm not advocating a change: I'll leave it to Karl to do that if he
really wants to and to the maintainers to decide whether it's worth
doing. But it can be done (more or less). And maybe somebody will come
up with a better way than the proof-of-concept that I'm attaching here.

Nick

PS. BTW, if you look at the attachment, it helps if you have a wide window,
    something like 165 characters wide.


[-- Attachment #2: Karl's folly --]
[-- Type: text/plain, Size: 3016 bytes --]

* Modify org-ts-regexpr0 to satisfy Karl Voit :-)

This is an attempt to allow timestamps without a day-of-week.

Modifying complicated regexps is no fun: I had to resort to re-builder
to figure out what was going on. The problem was that the first " *" which
is supposed to eat all the spaces between the "2012-03-05" part and
the "Mon" part, would, in the absence of the "Mon" part, eat the space
before the time spec "08:00". But the regexp for the time part explicitly
requires that space (see the second sp in the explanatory line below).
So I tried making the DOW part optional by appending a ?, but that didn't
do the trick. re-builder showed the space problem: there was no space left
to match if DOW is absent, so I changed the literal space " " to 0 or more
spaces " *" and that did the trick (with some caveats noted below).

In the code block below, tslist is a list of test cases. The code
block sets org-ts-regexp0 first to its original value, then to the
changed value (the first one is there just for comparison between the
two). I've marked the changes (of which, the second is the significant
one). We map the function org-parse-time-string over the test
cases and get a list of results. The results look reasonable, but the
new regexp allows funny-looking input like the second and fourth test
case.

One additional caveat is that org-odt uses org-ts-regexp0, so changing
it without checking org-odt's use is fraught with peril, but that's what
I've done: I have no idea whether it causes problems in org-odt.

#+BEGIN_SRC elisp
  ;                           year_______    -   month______    -   day________    sp   DOW____________       sp  hour_________    :   minutes____
  (setq org-ts-regexp0 "\\(\\([0-9]\\{4\\}\\)-\\([0-9]\\{2\\}\\)-\\([0-9]\\{2\\}\\) *\\([^]+0-9>\r\n -]*\\)\\( \\([0-9]\\{1,2\\}\\):\\([0-9]\\{2\\}\\)\\)?\\)")
  ; changed                                                                                                v    v
  (setq org-ts-regexp0 "\\(\\([0-9]\\{4\\}\\)-\\([0-9]\\{2\\}\\)-\\([0-9]\\{2\\}\\) *\\([^]+0-9>\r\n -]*\\)?\\( *\\([0-9]\\{1,2\\}\\):\\([0-9]\\{2\\}\\)\\)?\\)")
  (setq tslist '(
          "<2012-03-05 Mon 08:00>"
          "<2012-03-05 Mon08:00>"
          "<2012-03-05 Mon 8:00>"
          "<2012-03-05 Mon8:00>"
          "<2012-03-05 Mon     8:00>"
          "<2012-03-05    Mon     8:00>"
          "<2012-03-05 Mon >"
          "<2012-03-05 Mon>"
          "<2012-03-05 08:00>"
          "<2012-03-05 8:00>"
          "<2012-03-05       8:00>"
          "<2012-03-05>"))
  
  (pp (mapcar 'org-parse-time-string tslist))
  
  
#+END_SRC

#+RESULTS:
#+begin_example
((0 0 8 5 3 2012 nil nil nil)
 (0 0 8 5 3 2012 nil nil nil)
 (0 0 8 5 3 2012 nil nil nil)
 (0 0 8 5 3 2012 nil nil nil)
 (0 0 8 5 3 2012 nil nil nil)
 (0 0 8 5 3 2012 nil nil nil)
 (0 0 0 5 3 2012 nil nil nil)
 (0 0 0 5 3 2012 nil nil nil)
 (0 0 8 5 3 2012 nil nil nil)
 (0 0 8 5 3 2012 nil nil nil)
 (0 0 8 5 3 2012 nil nil nil)
 (0 0 0 5 3 2012 nil nil nil))
#+end_example

Re: [bug] org-agenda-write does not handle date stamps without day of week

[Carsten Dominik]

[-- Attachment #1: Type: text/plain, Size: 4060 bytes --]


On 17.3.2012, at 07:45, Nick Dokos wrote:

> Karl Voit <devnull@Karl-Voit.at> wrote:
> 
>> * Nick Dokos <nicholas.dokos@hp.com> wrote:
>>> Karl Voit <devnull@Karl-Voit.at> wrote:
>>> 
>>> For me, it was a "no time to work on org - stash it"...
>> 
>> OK. I just wanted to make sure that it *is* on someone's todo list
>> :-)
>> 
>>>> * Karl Voit <devnull@Karl-Voit.at> wrote:
>>>>> 
>>>>> * <2012-03-05 08:00-09:00> Wrong: ends up as full day event
>>> 
>>> org-agenda-write calls org-export-icalendar which calls org-print-icalendar-entries
>>> which loops over all the entries and parses them, decomposing them into timestamps.
>>> Each timestamp is then passed to org-parse-time-string. It's this one that cannot
>>> handle non-standard formats: it uses a regexp and assumes that all the matched parts
>>> are going to be in fixed places:
>>> 
>>> As to how to fix it, there are several possibilities:
>>> 
>>> 1. fix your scripts that produce time stamps to include day-of-week.
>> 
>> Sorry, deriving DOW from an arbitrary timestamps from arbitrary data
>> sources is either pretty time consuming (calendar calculations) or
>> simply hard to calculate.
>> 
>> Outside Org-mode, DOW is seldom part of time-stamp data :-(
>> 
>>> 2. change the callers of org-parse-time-string to make sure that DOW is included.
>>> There are roughly three dozen callers, so 2. is possible but a pain.
>> 
>> Ack.
>> 
>>> 3. change just one caller: org-print-icalendar-entries to make sure that DOW is included.
>>> 3. is simple but ugly as sin, 
>> 
>> Ouch, ack :-)
>> 
>>> 4. change org-parse-time-string to handle a missing DOW.
>>> 4. is the best way to handle it within org.
>> 
>> Full ack.
>> 
>>> I vote for 1. where *you* have to do all the work ;-)
>> 
>> YMMD :-)
>> 
>> If my brain would be compatible to ELISP, I'd send a patch.
>> Promised.
>> 
>> But I'll take my chance and wait for someone else (you?)
>> implementing 4. to resolve this issue for everybody. I really
>> appreciate every second you guys invest in maintaining Org-mode!
>> 
> 
> I don't know about you, but whenever I engage in hand-to-hand combat
> with a complicated regexp, I come out bruised, muddied and a lot worse
> for wear. In any case, I'm attaching an org file with my investigations.
> It contains a description and a code block for testing.
> 
> I hope that the attachment will come through unscathed: it contains
> regexps, and munging a regexp that looks like hen scratchings in the
> first place through uncooperative mailers is not something to be
> relished.
> 
> BTW, I'm not advocating a change: I'll leave it to Karl to do that if he
> really wants to and to the maintainers to decide whether it's worth
> doing. But it can be done (more or less). And maybe somebody will come
> up with a better way than the proof-of-concept that I'm attaching here.

Hi Nick and Karl,

since we did make a change to Org a while ago to allow date stamps
without the name of the day, I think it is only consequent to also do
it for this case.  Must have slipped our attention back then.
The only thing we must ensure is that this regexp matches fast
as it is used a lot.

Nick's proposal works, except for the fact that is also matches when
the time is directly attached to day name.  Maybe it is cleaner to
not match in this case.

If we are going to make the day name optional, then it is better to
include matching of the whitespace after the date into the day-name
part of the regexp.

I am attaching Nick's file again, with a third proposal for an
updated regexp.

As far as speed is concerned, this regexp will, if there is name
and time, match directly and straight.  If the date name is
missing, it will notice on the first digit belonging to
the time and switch without backtracking (well, minimal
backtracking when there are multiple spaces) to the regexp
section for the time of day.  Rematching the spaces after the
date will be the only overhead.

Cheers

- Carsten


[-- Attachment #2: reproducer.org --]
[-- Type: application/octet-stream, Size: 3291 bytes --]

* Modify org-ts-regexpr0 to satisfy Karl Voit :-)

This is an attempt to allow timestamps without a day-of-week.

Modifying complicated regexps is no fun: I had to resort to re-builder
to figure out what was going on. The problem was that the first " *" which
is supposed to eat all the spaces between the "2012-03-05" part and
the "Mon" part, would, in the absence of the "Mon" part, eat the space
before the time spec "08:00". But the regexp for the time part explicitly
requires that space (see the second sp in the explanatory line below).
So I tried making the DOW part optional by appending a ?, but that didn't
do the trick. re-builder showed the space problem: there was no space left
to match if DOW is absent, so I changed the literal space " " to 0 or more
spaces " *" and that did the trick (with some caveats noted below).

In the code block below, tslist is a list of test cases. The code
block sets org-ts-regexp0 first to its original value, then to the
changed value (the first one is there just for comparison between the
two). I've marked the changes (of which, the second is the significant
one). We map the function org-parse-time-string over the test
cases and get a list of results. The results look reasonable, but the
new regexp allows funny-looking input like the second and fourth test
case.

One additional caveat is that org-odt uses org-ts-regexp0, so changing
it without checking org-odt's use is fraught with peril, but that's what
I've done: I have no idea whether it causes problems in org-odt.

#+BEGIN_SRC elisp
  ;                           year_______    -   month______    -   day________    sp   DOW____________       sp  hour_________    :   minutes____
  (setq org-ts-regexp0 "\\(\\([0-9]\\{4\\}\\)-\\([0-9]\\{2\\}\\)-\\([0-9]\\{2\\}\\) *\\([^]+0-9>\r\n -]*\\)\\( \\([0-9]\\{1,2\\}\\):\\([0-9]\\{2\\}\\)\\)?\\)")
  ; changed                                                                                                v    v
  (setq org-ts-regexp0 "\\(\\([0-9]\\{4\\}\\)-\\([0-9]\\{2\\}\\)-\\([0-9]\\{2\\}\\) *\\([^]+0-9>\r\n -]*\\)?\\( *\\([0-9]\\{1,2\\}\\):\\([0-9]\\{2\\}\\)\\)?\\)")
  ; New version by Carsten                                                         ^^vv                         v
  (setq org-ts-regexp0 "\\(\\([0-9]\\{4\\}\\)-\\([0-9]\\{2\\}\\)-\\([0-9]\\{2\\}\\)\\( +[^]+0-9>\r\n -]+\\)?\\( +\\([0-9]\\{1,2\\}\\):\\([0-9]\\{2\\}\\)\\)?\\)"
  (setq tslist '(
          "<2012-03-05 Mon 08:00>"
          "<2012-03-05 Mon08:00>"
          "<2012-03-05 Mon 8:00>"
          "<2012-03-05 Mon8:00>"
          "<2012-03-05 Mon     8:00>"
          "<2012-03-05    Mon     8:00>"
          "<2012-03-05 Mon >"
          "<2012-03-05 Mon>"
          "<2012-03-05 08:00>"
          "<2012-03-05 8:00>"
          "<2012-03-05       8:00>"
          "<2012-03-05>"))
  
  (pp (mapcar 'org-parse-time-string tslist))
  
  
#+END_SRC

#+RESULTS:
#+begin_example
((0 0 8 5 3 2012 nil nil nil)
 (0 0 8 5 3 2012 nil nil nil)
 (0 0 8 5 3 2012 nil nil nil)
 (0 0 8 5 3 2012 nil nil nil)
 (0 0 8 5 3 2012 nil nil nil)
 (0 0 8 5 3 2012 nil nil nil)
 (0 0 0 5 3 2012 nil nil nil)
 (0 0 0 5 3 2012 nil nil nil)
 (0 0 8 5 3 2012 nil nil nil)
 (0 0 8 5 3 2012 nil nil nil)
 (0 0 8 5 3 2012 nil nil nil)
 (0 0 0 5 3 2012 nil nil nil))
#+end_example


[-- Attachment #3: Type: text/plain, Size: 159 bytes --]





> 
> Nick
> 
> PS. BTW, if you look at the attachment, it helps if you have a wide window,
>    something like 165 characters wide.
> 
> <reproducer.org>

Re: [bug] org-agenda-write does not handle date stamps without day of week

[Carsten Dominik]

I have just pushed the new version of this regexp into master.

Karl, let me know if it fixes the issues you where having
with writing agendas.

- Carsten

On 19.3.2012, at 10:12, Carsten Dominik wrote:

> 
> On 17.3.2012, at 07:45, Nick Dokos wrote:
> 
>> Karl Voit <devnull@Karl-Voit.at> wrote:
>> 
>>> * Nick Dokos <nicholas.dokos@hp.com> wrote:
>>>> Karl Voit <devnull@Karl-Voit.at> wrote:
>>>> 
>>>> For me, it was a "no time to work on org - stash it"...
>>> 
>>> OK. I just wanted to make sure that it *is* on someone's todo list
>>> :-)
>>> 
>>>>> * Karl Voit <devnull@Karl-Voit.at> wrote:
>>>>>> 
>>>>>> * <2012-03-05 08:00-09:00> Wrong: ends up as full day event
>>>> 
>>>> org-agenda-write calls org-export-icalendar which calls org-print-icalendar-entries
>>>> which loops over all the entries and parses them, decomposing them into timestamps.
>>>> Each timestamp is then passed to org-parse-time-string. It's this one that cannot
>>>> handle non-standard formats: it uses a regexp and assumes that all the matched parts
>>>> are going to be in fixed places:
>>>> 
>>>> As to how to fix it, there are several possibilities:
>>>> 
>>>> 1. fix your scripts that produce time stamps to include day-of-week.
>>> 
>>> Sorry, deriving DOW from an arbitrary timestamps from arbitrary data
>>> sources is either pretty time consuming (calendar calculations) or
>>> simply hard to calculate.
>>> 
>>> Outside Org-mode, DOW is seldom part of time-stamp data :-(
>>> 
>>>> 2. change the callers of org-parse-time-string to make sure that DOW is included.
>>>> There are roughly three dozen callers, so 2. is possible but a pain.
>>> 
>>> Ack.
>>> 
>>>> 3. change just one caller: org-print-icalendar-entries to make sure that DOW is included.
>>>> 3. is simple but ugly as sin, 
>>> 
>>> Ouch, ack :-)
>>> 
>>>> 4. change org-parse-time-string to handle a missing DOW.
>>>> 4. is the best way to handle it within org.
>>> 
>>> Full ack.
>>> 
>>>> I vote for 1. where *you* have to do all the work ;-)
>>> 
>>> YMMD :-)
>>> 
>>> If my brain would be compatible to ELISP, I'd send a patch.
>>> Promised.
>>> 
>>> But I'll take my chance and wait for someone else (you?)
>>> implementing 4. to resolve this issue for everybody. I really
>>> appreciate every second you guys invest in maintaining Org-mode!
>>> 
>> 
>> I don't know about you, but whenever I engage in hand-to-hand combat
>> with a complicated regexp, I come out bruised, muddied and a lot worse
>> for wear. In any case, I'm attaching an org file with my investigations.
>> It contains a description and a code block for testing.
>> 
>> I hope that the attachment will come through unscathed: it contains
>> regexps, and munging a regexp that looks like hen scratchings in the
>> first place through uncooperative mailers is not something to be
>> relished.
>> 
>> BTW, I'm not advocating a change: I'll leave it to Karl to do that if he
>> really wants to and to the maintainers to decide whether it's worth
>> doing. But it can be done (more or less). And maybe somebody will come
>> up with a better way than the proof-of-concept that I'm attaching here.
> 
> Hi Nick and Karl,
> 
> since we did make a change to Org a while ago to allow date stamps
> without the name of the day, I think it is only consequent to also do
> it for this case.  Must have slipped our attention back then.
> The only thing we must ensure is that this regexp matches fast
> as it is used a lot.
> 
> Nick's proposal works, except for the fact that is also matches when
> the time is directly attached to day name.  Maybe it is cleaner to
> not match in this case.
> 
> If we are going to make the day name optional, then it is better to
> include matching of the whitespace after the date into the day-name
> part of the regexp.
> 
> I am attaching Nick's file again, with a third proposal for an
> updated regexp.
> 
> As far as speed is concerned, this regexp will, if there is name
> and time, match directly and straight.  If the date name is
> missing, it will notice on the first digit belonging to
> the time and switch without backtracking (well, minimal
> backtracking when there are multiple spaces) to the regexp
> section for the time of day.  Rematching the spaces after the
> date will be the only overhead.
> 
> Cheers
> 
> - Carsten
> 
> <reproducer.org>
> 
> 
> 
>> 
>> Nick
>> 
>> PS. BTW, if you look at the attachment, it helps if you have a wide window,
>>   something like 165 characters wide.
>> 
>> <reproducer.org>
> 

Re: [bug] org-agenda-write does not handle date stamps without day of week

[Karl Voit]

* Carsten Dominik <carsten.dominik@gmail.com> wrote:
> I have just pushed the new version of this regexp into master.

Very cool!

> Karl, let me know if it fixes the issues you where having
> with writing agendas.

My agenda looks fine. But generating the ics file still results in
events without the time information:

,----[ my function generating the ics ]
| (defun vk-export-agenda()
|   "Exports monthly Org-mode agenda to agenda.ics file"
|   (interactive)
|   (org-agenda-list nil nil 60)
|   (org-agenda-write "~/share/all/org-mode/agenda.ics")
| )
`----

with:

,----[ example events ]
| ** <2012-03-20 8:00-9:30> Test
| :PROPERTIES:
| :CREATED: [2012-03-20 Tue 10:38]
| :END:
|
| ** <2012-03-20 8:12-13:12> Test
| :PROPERTIES:
| :CREATED: [2012-03-20 Tue 10:40]
| :END:
`----

results in:

,----[ ics result ]
| vk@gary ~all/org-mode (git)-[master] % grep -C 2 Test agenda.ics
| DTSTART;VALUE=DATE:20120320
| DTEND;VALUE=DATE:20120320
| SUMMARY: Test
| CATEGORIES:misc
| END:VEVENT
| --
| DTSTART;VALUE=DATE:20120320
| DTEND;VALUE=DATE:20120320
| SUMMARY: Test
| CATEGORIES:misc
| END:VEVENT
| vk@gary ~all/org-mode (git)-[master] %
`----

-- 
Karl Voit

Re: [bug] org-agenda-write does not handle date stamps without day of week

[Carsten Dominik]

Hi Karl,

I am not able to reproduce this.  I am getting

BEGIN:VEVENT
UID: TS-13B9F6A7-BF2B-445C-84CA-7FB8ED15E0C1
DTSTART:20120320T080000
DTEND:20120320T093000
SUMMARY: Test
CATEGORIES:x
END:VEVENT
BEGIN:VEVENT
UID: TS-AF0458B6-C5D6-4419-9DBB-FAB0651EF870
DTSTART:20120320T081200
DTEND:20120320T131200
SUMMARY: Test
CATEGORIES:x
END:VEVENT

Looks right to me.

- Carsten

On 20.3.2012, at 13:26, Karl Voit wrote:

> * Carsten Dominik <carsten.dominik@gmail.com> wrote:
>> I have just pushed the new version of this regexp into master.
> 
> Very cool!
> 
>> Karl, let me know if it fixes the issues you where having
>> with writing agendas.
> 
> My agenda looks fine. But generating the ics file still results in
> events without the time information:
> 
> ,----[ my function generating the ics ]
> | (defun vk-export-agenda()
> |   "Exports monthly Org-mode agenda to agenda.ics file"
> |   (interactive)
> |   (org-agenda-list nil nil 60)
> |   (org-agenda-write "~/share/all/org-mode/agenda.ics")
> | )
> `----
> 
> with:
> 
> ,----[ example events ]
> | ** <2012-03-20 8:00-9:30> Test
> | :PROPERTIES:
> | :CREATED: [2012-03-20 Tue 10:38]
> | :END:
> |
> | ** <2012-03-20 8:12-13:12> Test
> | :PROPERTIES:
> | :CREATED: [2012-03-20 Tue 10:40]
> | :END:
> `----
> 
> results in:
> 
> ,----[ ics result ]
> | vk@gary ~all/org-mode (git)-[master] % grep -C 2 Test agenda.ics
> | DTSTART;VALUE=DATE:20120320
> | DTEND;VALUE=DATE:20120320
> | SUMMARY: Test
> | CATEGORIES:misc
> | END:VEVENT
> | --
> | DTSTART;VALUE=DATE:20120320
> | DTEND;VALUE=DATE:20120320
> | SUMMARY: Test
> | CATEGORIES:misc
> | END:VEVENT
> | vk@gary ~all/org-mode (git)-[master] %
> `----
> 
> -- 
> Karl Voit
> 
> 

Re: [bug] org-agenda-write does not handle date stamps without day of week

[Bastien]

Hi Carsten and all,

Carsten Dominik <carsten.dominik@gmail.com> writes:

> since we did make a change to Org a while ago to allow date stamps
> without the name of the day, I think it is only consequent to also do
> it for this case.  Must have slipped our attention back then.
> The only thing we must ensure is that this regexp matches fast
> as it is used a lot.
>
> Nick's proposal works, except for the fact that is also matches when
> the time is directly attached to day name.  Maybe it is cleaner to
> not match in this case.
>
> If we are going to make the day name optional, then it is better to
> include matching of the whitespace after the date into the day-name
> part of the regexp.

Agreed.

> I am attaching Nick's file again, with a third proposal for an
> updated regexp.

I tested it from reproducer.org (had to add a missing parenthesis) 
and it seems to work fine.

Carsten, please feel free to apply a patch with your new org-ts-regexp0
proposal against master HEAD.

Thanks!

-- 
 Bastien

Re: [bug] org-agenda-write does not handle date stamps without day of week

[Nick Dokos]

Bastien <bzg@gnu.org> wrote:

> Hi Carsten and all,
> 
> Carsten Dominik <carsten.dominik@gmail.com> writes:
> 
> > since we did make a change to Org a while ago to allow date stamps
> > without the name of the day, I think it is only consequent to also do
> > it for this case.  Must have slipped our attention back then.
> > The only thing we must ensure is that this regexp matches fast
> > as it is used a lot.
> >
> > Nick's proposal works, except for the fact that is also matches when
> > the time is directly attached to day name.  Maybe it is cleaner to
> > not match in this case.
> >
> > If we are going to make the day name optional, then it is better to
> > include matching of the whitespace after the date into the day-name
> > part of the regexp.
> 
> Agreed.
> 
> > I am attaching Nick's file again, with a third proposal for an
> > updated regexp.
> 
> I tested it from reproducer.org (had to add a missing parenthesis) 
> and it seems to work fine.
> 
> Carsten, please feel free to apply a patch with your new org-ts-regexp0
> proposal against master HEAD.
> 

I think he did already (but please check!):

commit 7d84b300f4cba8b726662113d4ce4ff40a76c5b5

Thanks,
Nick

Re: [bug] org-agenda-write does not handle date stamps without day of week

[Bastien]

Nick Dokos <nicholas.dokos@hp.com> writes:

> I think he did already (but please check!):
>
> commit 7d84b300f4cba8b726662113d4ce4ff40a76c5b5

You're right -- time to power off my computer :)

Thanks!

-- 
 Bastien

Re: [bug] org-agenda-write does not handle date stamps without day of week

[Nick Dokos]

Karl Voit <devnull@Karl-Voit.at> wrote:

> * Carsten Dominik <carsten.dominik@gmail.com> wrote:
> > I have just pushed the new version of this regexp into master.
> 
> Very cool!
> 
> > Karl, let me know if it fixes the issues you where having
> > with writing agendas.
> 
> My agenda looks fine. But generating the ics file still results in
> events without the time information:
> 
> ,----[ my function generating the ics ]
> | (defun vk-export-agenda()
> |   "Exports monthly Org-mode agenda to agenda.ics file"
> |   (interactive)
> |   (org-agenda-list nil nil 60)
> |   (org-agenda-write "~/share/all/org-mode/agenda.ics")
> | )
> `----
> 
> with:
> 
> ,----[ example events ]
> | ** <2012-03-20 8:00-9:30> Test
> | :PROPERTIES:
> | :CREATED: [2012-03-20 Tue 10:38]
> | :END:
> |
> | ** <2012-03-20 8:12-13:12> Test
> | :PROPERTIES:
> | :CREATED: [2012-03-20 Tue 10:40]
> | :END:
> `----
> 
> results in:
> 
> ,----[ ics result ]
> | vk@gary ~all/org-mode (git)-[master] % grep -C 2 Test agenda.ics
> | DTSTART;VALUE=DATE:20120320
> | DTEND;VALUE=DATE:20120320
> | SUMMARY: Test
> | CATEGORIES:misc
> | END:VEVENT
> | --
> | DTSTART;VALUE=DATE:20120320
> | DTEND;VALUE=DATE:20120320
> | SUMMARY: Test
> | CATEGORIES:misc
> | END:VEVENT
> | vk@gary ~all/org-mode (git)-[master] %
> `----
> 

I get

BEGIN:VEVENT
UID: TS-11c60341-06a6-4b6f-bb22-2ccd739bbdc8
DTSTART:20120320T080000
DTEND:20120320T093000
SUMMARY: Test
CATEGORIES:foo
END:VEVENT

BEGIN:VEVENT
UID: TS-dc99f34e-e24a-4ecc-add5-fa67649b005e
DTSTART:20120320T081200
DTEND:20120320T131200
SUMMARY: Test
CATEGORIES:foo
END:VEVENT

which looks correct to me.

Check that you picked up Carsten's fix (commit ):

      git show 7d84b300f4cba8b726662113d4ce4ff40a76c5b5

If you did M-x org-reload RET only, then that's probably the reason:
check to see if the value of org-ts-regexp0 has been updated; in my case
it hadn't and I ended up doing M-x load-file RET /path/to/org.el RET.
Alternatively, restart emacs.

Nick

Re: [bug] org-agenda-write does not handle date stamps without day of week

[Karl Voit]

Hi!

* Nick Dokos <nicholas.dokos@hp.com> wrote:
>
> Check that you picked up Carsten's fix (commit ):
>       git show 7d84b300f4cba8b726662113d4ce4ff40a76c5b5

I do. I did a «git pull» and the «git show» from above shows me
Casten's fix.

> If you did M-x org-reload RET only, then that's probably the reason:
> check to see if the value of org-ts-regexp0 has been updated; in my case
> it hadn't and I ended up doing M-x load-file RET /path/to/org.el RET.
> Alternatively, restart emacs.

Being very unsure what needs to be restarted at such updates, I
quit Emacs before the git commands and start it again afterwards.

So I quit Emacs again, did another «git pull» a couple of minutes
ago, compiled all *.el once again, started Emacs, generated the ICS
file and now the times get written to ICS:

,----[ Org-mode entries ]
| ** <2012-03-21 08:12> TSTest
| :PROPERTIES:
| :CREATED: [2012-03-20 Tue 10:38]
| :END:
|
| ** <2012-03-21 08:07-13:12> TSTest
| :PROPERTIES:
| :CREATED: [2012-03-20 Tue 10:40]
| :END:
`----

,----[ ICS entries ]
| BEGIN:VEVENT
| UID: TS-4ae514e8-210c-408b-bb08-254dcb3a5277
| DTSTART:20120321T081200
| DTEND:20120321T101200
| SUMMARY: TSTest
| CATEGORIES:misc
| END:VEVENT
| BEGIN:VEVENT
| UID: TS-cab4b928-1960-4335-b2e8-2179e66eec5a
| DTSTART:20120321T080700
| DTEND:20120321T131200
| SUMMARY: TSTest
| CATEGORIES:misc
| END:VEVENT
`----

ICS look great! Thank you guys for tracing the problem and fixing
it! I really appreciate the help!


So far, Google calendar does not use the time information of those
two test entries :-(

But since the ICS data seems to be OK there might be another hiccup
by Google this time. I created new test entries with different names
and wait for Google calendar to re-read the new data.

I am optimistic that this is just a caching problem of Google
because I did not change the descriptions during my last test.

-- 
Karl Voit

Re: [bug] org-agenda-write does not handle date stamps without day of week

[Carsten Dominik]

On 21.3.2012, at 17:32, Karl Voit wrote:

> Hi!
> 
> * Nick Dokos <nicholas.dokos@hp.com> wrote:
>> 
>> Check that you picked up Carsten's fix (commit ):
>>      git show 7d84b300f4cba8b726662113d4ce4ff40a76c5b5
> 
> I do. I did a «git pull» and the «git show» from above shows me
> Casten's fix.
> 
>> If you did M-x org-reload RET only, then that's probably the reason:
>> check to see if the value of org-ts-regexp0 has been updated; in my case
>> it hadn't and I ended up doing M-x load-file RET /path/to/org.el RET.
>> Alternatively, restart emacs.
> 
> Being very unsure what needs to be restarted at such updates,

The reason why just reloading a file does not work here is that the
regexp is defined as a constant (defconst), which will not be changed on
executing the defconst again.  So exiting and restarting Emacs is the
right thing to do.

- Carsten

> I
> quit Emacs before the git commands and start it again afterwards.
> 
> So I quit Emacs again, did another «git pull» a couple of minutes
> ago, compiled all *.el once again, started Emacs, generated the ICS
> file and now the times get written to ICS:
> 
> ,----[ Org-mode entries ]
> | ** <2012-03-21 08:12> TSTest
> | :PROPERTIES:
> | :CREATED: [2012-03-20 Tue 10:38]
> | :END:
> |
> | ** <2012-03-21 08:07-13:12> TSTest
> | :PROPERTIES:
> | :CREATED: [2012-03-20 Tue 10:40]
> | :END:
> `----
> 
> ,----[ ICS entries ]
> | BEGIN:VEVENT
> | UID: TS-4ae514e8-210c-408b-bb08-254dcb3a5277
> | DTSTART:20120321T081200
> | DTEND:20120321T101200
> | SUMMARY: TSTest
> | CATEGORIES:misc
> | END:VEVENT
> | BEGIN:VEVENT
> | UID: TS-cab4b928-1960-4335-b2e8-2179e66eec5a
> | DTSTART:20120321T080700
> | DTEND:20120321T131200
> | SUMMARY: TSTest
> | CATEGORIES:misc
> | END:VEVENT
> `----
> 
> ICS look great! Thank you guys for tracing the problem and fixing
> it! I really appreciate the help!
> 
> 
> So far, Google calendar does not use the time information of those
> two test entries :-(
> 
> But since the ICS data seems to be OK there might be another hiccup
> by Google this time. I created new test entries with different names
> and wait for Google calendar to re-read the new data.
> 
> I am optimistic that this is just a caching problem of Google
> because I did not change the descriptions during my last test.
> 
> -- 
> Karl Voit
> 
> 

Re: Plotting (with gnuplot) using dates timestamps

[Nick Dokos]

Karl Voit <devnull@Karl-Voit.at> wrote:

> * Alan Schmitt <alan.schmitt@polytechnique.org> wrote:
> >
> > Thanks a lot for all the suggestions. The part I was missing was the preprocessing from babel. Using simply
> >
> > set xdata time
> > set timefmt "%Y-%m-%d"
> >
> > did the trick.
> 
> I just wanted to check this out but I am not able to produce the
> plot :-( I just get the coordinate system but with an empty plane
> (no points).
> 
> Can you please post the while working example (or correct my error)?
> 

Did you try the version I posted? I believe it's a complete working
example.

Nick

> Thanks!
> 
> 
> #+tblname: data-table
> | Date             |     HP |     HC |
> |------------------+--------+--------|
> | [2011-08-20 Sat] | 006815 | 008399 |
> | [2011-08-29 Mon] | 006840 | 008438 |
> | [2011-09-11 Sun] | 006946 | 008552 |
> | [2011-12-11 Sun] | 007805 | 009603 |
> | [2012-03-04 Sun] | 008800 | 010826 |
> | [2012-03-11 Sun] | 008876 | 010930 |
> | [2012-03-25 Sun] | 009015 | 011121 |
> 
> #+begin_src gnuplot :var data=data-table
> reset
> 
> set title "Consommation Electrique"
> set size ratio square
> 
> set xlabel "Date"
> set xdata time
> set timefmt "[%Y-%m-%d %a]"
> set xrange ["[2011-08-01 foo]":"[2012-04-01 bar]"]
> 
> set yrange [6800:9100]
> set ylabel "HP"
> set ytics nomirror
> 
> set y2range [8000:12000]
> set y2label "HC"
> set y2tics nomirror
> 
> set style data points
> plot data using 1:2 axis x1y1 title 'HP', \
>      data using 1:3 axis x1y2 title 'HC'
> #+end_src
> 
> 
> -- 
> Karl Voit
> 
> Hallo, mein Name ist Web 2.0. Gib mir dein Adressbuch und lass mich Emails lesen
> 
> 

Re: table: referencing row of other table

[Nick Dokos]

Karl Voit <devnull@Karl-Voit.at> wrote:

> Hi!
> 
> I want to use the column "average" of the first table to fill the
> column "h1" in the second one.
> 
> #+TBLNAME: 2012-08-12vkmeasure
> | tags/item |    m1 |    m2 |    m3 | average |
> |-----------+-------+-------+-------+---------|
> |         4 |  0.02 |  0.03 |  0.02 |    0.02 |
> |         5 |  0.06 |  0.12 |  0.06 |    0.08 |
> |         6 |  0.31 |  0.53 |  0.29 |    0.38 |
> |         7 |  3.83 |  4.08 |  4.48 |    4.13 |
> |         8 | 85.33 | 89.22 | 92.07 |   88.87 |
> #+TBLFM: $5 = vmean($2..$4);%.2f
> 
> | tags/item | h1 [s] | h2 [s] |
> |-----------+--------+--------|
> |         4 |        |        |
> |         5 |        |        |
> |         6 |        |        |
> |         7 |        |        |
> |         8 |        |        |
> #+TBLFM: <fillme>
> 

#+TBLFM: $2 = remote(2012-08-12vkmeasure, @@#$5)

Nick

> I tried following references but failed so far:
> 
> @2$2..@6$2=subscr(remote(2012-08-12vkmeasure,@2$5..@6$5),$#) #-> 0.08 0.08 ...
> @2$2=remote(2012-08-12vkmeasure,@2$5) -> works for one value
> @2$2..@6$2=subscr(remote(2012-08-12vkmeasure,@2$5..@6$5),$#) #-> 0.08 0.08 ...
> $2=subscr(remote(2012-08-12vkmeasure,@I$>..@II$>),$#) 
> 
> What did I do wrong?
> 
> How is the correct TBLFM in order to get 0.02 ... 88.87 into the
> h1-column?
> 
> Thanks for your help!
> 
> -- 
> Karl Voit
> 
> 

Re: table: referencing row of other table

[Karl Voit]

* Nick Dokos <nicholas.dokos@hp.com> wrote:
> Karl Voit <devnull@Karl-Voit.at> wrote:
>
>> I want to use the column "average" of the first table to fill the
>> column "h1" in the second one.
>
> #+TBLFM: $2 = remote(2012-08-12vkmeasure, @@#$5)

Thanks!

... and now I also found the corresponding documentation which I
overlooked twice :-)

-- 
Karl Voit

Re: What HW/SW issues influence org/Emacs performance?

[Nick Dokos]

Karl Voit <devnull@Karl-Voit.at> wrote:

> 
> What do you think are the most important factors?
> 
> - CPU speed (of one core)
> - type of CPU
> - # CPU cores
> - 32bit kernel vs. 64bit kernel
> - faster I/O (through HDD->SSD or even faster SSD)
> - choice of the file system (optimized for something? what?)
> - operating system 
> - possible ELISP optimizations (are there any besides compiling?)
> 

How much memory do you have? Are you swapping at all? If yes, get more
memory.

Nick

Re: What HW/SW issues influence org/Emacs performance?

[Karl Voit]

* Nick Dokos <nicholas.dokos@hp.com> wrote:
> Karl Voit <devnull@Karl-Voit.at> wrote:
>
>> What do you think are the most important factors?
>> 
>> - CPU speed (of one core)
>> - type of CPU
>> - # CPU cores
>> - 32bit kernel vs. 64bit kernel
>> - faster I/O (through HDD->SSD or even faster SSD)
>> - choice of the file system (optimized for something? what?)
>> - operating system 
>> - possible ELISP optimizations (are there any besides compiling?)
>
> How much memory do you have? Are you swapping at all? If yes, get more
> memory.

Sorry - I forgot to mention: 4GB of RAM (but only 32bit Kernel). 

I am not running much memory-intense things. Therefore I do have
enough free RAM to avoid any swapping.

-- 
Karl Voit

Re: Replacement for org-make-link

[Nick Dokos]

Karl Voit <devnull@Karl-Voit.at> wrote:

> Hi!
> 
> http://orgmode.org/Changes.html tells me that org-make-link has been
> deleted.
> 
> I am using following lines to add links to references I manage using
> Org-mode: (from [1])
> 
> ,----
> | (defun org-ref-complete-link (&optional arg)
> |   "Create a reference link using completion."
> |   (let (file link)
> |        (setq file (read-file-name "ref: " "~/archive/papers_from_web/"))
> |        (let ((pwd (file-name-as-directory (expand-file-name ".")))
> |          (pwd1 (file-name-as-directory (abbreviate-file-name
> |                         (expand-file-name ".")))))
> |          (setq file (string-replace "~/archive/papers_from_web/" "" file))
> |          (setq file (string-replace pwd "" (string-replace pwd1 "" file)))
> |          (setq file (string-replace ".bib" "" file))
> |          (setq file (string-replace ".pdf" "" file))
> |          (setq link (org-make-link "ref:" file)))
> |     link))
> `----
> 
> Can somebody with more knowledge than me suggest, what I should do
> instead?
> 

Replace it with concat, I think. That's all that org-make-link used to do:

,----
| (defun org-make-link (&rest strings)
|   "Concatenate STRINGS."
|   (apply 'concat strings))
`----

In fact the commit says:

,----
| commit 96551f3dd8885dee6972c70ce06888fd3d5f4dd4
| Author: Bastien Guerry <bzg@altern.org>
| Date:   Fri Aug 3 19:03:21 2012 +0200
| 
|     Delete `org-make-link' and replace previous occurrences by `concat'.
`----

Nick

Re: Replacement for org-make-link

[Bastien]

Nick Dokos <nicholas.dokos@hp.com> writes:

> Replace it with concat, I think.

Yes, exactly.

-- 
 Bastien

Re: Replacement for org-make-link

[Karl Voit]

* Bastien <bzg@altern.org> wrote:
> Nick Dokos <nicholas.dokos@hp.com> writes:
>
>> Replace it with concat, I think.
>
> Yes, exactly.

Works like charm. Thanks!

-- 
Karl Voit

Re: comma as decimal point: bug?

[Nick Dokos]

Karl Voit <devnull@Karl-Voit.at> wrote:

> * Nicolas Goaziou <n.goaziou@gmail.com> wrote:
> >
> > Rainer Thiel <r.thiel@uni-jena.de> writes:
> >
> >> I have seen that as of v. 7.9, org tables are supposed to accept
> >> commas as decimal points which is useful if you set up tables to be
> >> used in in files where continental conventions are expected to be
> >> followed.
> >
> > Actually, this more subtle. Tables recognize numbers with a comma as the
> > decimal mark. But it only means such numbers will be properly aligned to
> > the right. Internally, calculations still require dot separator.
> 
> When I read the change-list of Org, I bet that there *will* be
> problems because of this cosmetic change :-)
> 
> From user experience perspective, it is quite clear that «Org
> recognizes numbers with commas» will be mixed up with «Org can
> calculate with numbers containing commas». You can not make this
> clear - even with renaming the settings variable.
> 
> Although I totally understand that numbers with commas should be
> made possible (I am a German speaking guy) I guess there will be
> even more issues in the future as long as cosmetics and
> functionality is not aligned.
> 

Agreed, but IIUC the main problem is that org depends on calc for
the calculations and calc is firmly in the "decimal point MUST be
a period" camp. So the cosmetics/functionality alignment would involve
surgery on calc, something that nobody should undertake lightly -
if at all :-).

I believe the above statement is correct but if anybody has better
information, please correct me.

Nick

Re: comma as decimal point: bug?

[Karl Voit]

* Nick Dokos <nicholas.dokos@hp.com> wrote:
> Karl Voit <devnull@Karl-Voit.at> wrote:
>
>> Although I totally understand that numbers with commas should be
>> made possible (I am a German speaking guy) I guess there will be
>> even more issues in the future as long as cosmetics and
>> functionality is not aligned.
>
> Agreed, but IIUC the main problem is that org depends on calc for
> the calculations and calc is firmly in the "decimal point MUST be
> a period" camp. 

Although I do not have much clue about «calc» I already assumed this
is the case.

> So the cosmetics/functionality alignment would involve surgery on
> calc, something that nobody should undertake lightly - if at all
> :-).

Totally agree.

The point is: new users and users that does not start reading
documentation before usage will not know this. And all I want to
mention is that although I myself (now) understand the reasons, I
also understand that (new) users will keep on misunderstand this
issue until the cosmetics is aligned with the functionality of
»calc«.

This results in lot of traffic on this list with bug-reports on
this.

To avoid this, normal users should not get the alignment per default
without having to read the explanation of the whole story.

This is a trade-off.

-- 
Karl Voit

Re: Wish: switch active time-stamps to inactive when CANCELED

[Nick Dokos]

Karl Voit <devnull@Karl-Voit.at> wrote:

> Hi!
> 
> I tend to keep events that I do not attend in my calendar for
> archiving reasons. But I switch their status to CANCELED.
> Unfortunately, those events still appear on my Google calendar and
> so forth. I get my usual reminders and it messes up my free/busy
> list.
> 
> Therefore, I am thinking that canceled events should turn their
> active time-stamps into inactive with an Org-mode option. This way,
> I can recall them in my agenda by typing ']' (show inactive
> time-stamps) but they do not interfere with my daily workflow.
> 
> What do you think of such an option for Org-mode?
> 

You can probably accomplish this with org-after-todo-state-change-hook:

Write a function that checks if org-state is "CANCELED", and if so looks
for active timestamps and calls org-toggle-timestamp-type on them.

Then add the function to the hook.

And don't forget to post the code here as well as on Worg :-)

Nick

Re: Wish: switch active time-stamps to inactive when CANCELED

[Karl Voit]

* Nick Dokos <nicholas.dokos@hp.com> wrote:
>
> You can probably accomplish this with org-after-todo-state-change-hook:
>
> Write a function that checks if org-state is "CANCELED", and if so looks
> for active timestamps and calls org-toggle-timestamp-type on them.
>
> Then add the function to the hook.

Oh, I would have guessed this algorithm overview/principle as well
:-)

But unfortunately, I know that little of ELISP, that I am not able
to implement it by myself :-(

-- 
Karl Voit

Re: Wish: switch active time-stamps to inactive when CANCELED

[Nick Dokos]

Karl Voit <devnull@Karl-Voit.at> wrote:

> * Nick Dokos <nicholas.dokos@hp.com> wrote:
> >
> > You can probably accomplish this with org-after-todo-state-change-hook:
> >
> > Write a function that checks if org-state is "CANCELED", and if so looks
> > for active timestamps and calls org-toggle-timestamp-type on them.
> >
> > Then add the function to the hook.
> 
> Oh, I would have guessed this algorithm overview/principle as well
> :-)
> 

It is more than an algorithm overview/principle: there are specific
suggestions in there - which hook to use, the state variable that you
need to check and the function to toggle the timestamp type are the
"real" thing: nothing vaporous about them.

> But unfortunately, I know that little of ELISP, that I am not able
> to implement it by myself :-(
> 

Time to learn some then - and there is no better way than scratching
your own itch :-)

Nick

Re: Wish: switch active time-stamps to inactive when CANCELED

[Karl Voit]

* Nick Dokos <nicholas.dokos@hp.com> wrote:
> Karl Voit <devnull@Karl-Voit.at> wrote:
>
>> Oh, I would have guessed this algorithm overview/principle as well
>> :-)
>
> It is more than an algorithm overview/principle: there are specific
> suggestions in there - which hook to use, the state variable that you
> need to check and the function to toggle the timestamp type are the
> "real" thing: nothing vaporous about them.

Agreed!

>> But unfortunately, I know that little of ELISP, that I am not able
>> to implement it by myself :-(
>
> Time to learn some then - and there is no better way than scratching
> your own itch :-)

Also: agreed! Problem is my workload and ... time.

ELISP is nothing you can learn in a single weekend. I've got the
feeling that this is a task that requires a longer period of time.
It's in SOMEDAY state in my Org along with:

- http://nic.ferrier.me.uk/blog/2012_07/tips-and-tricks-for-emacslisp
- http://cjohansen.no/an-introduction-to-elisp
- https://en.wikipedia.org/wiki/Elisp
- 58min [[https://www.youtube.com/watch?v%3DYMYehD9JD-Y&feature%3Drelated][LISP PROGRAMMING INTRODUCTION LECTURE ONE]]
- 9min [[https://www.youtube.com/watch?v%3DM-BFgErib4k&feature%3Drelated][An Brief Introduction to LISP - Pt 1, Syntax]] (1/4+)

-- 
Karl Voit

Re: Wish: switch active time-stamps to inactive when CANCELED

[Viktor Rosenfeld]

Hi,

Karl Voit wrote:

> ELISP is nothing you can learn in a single weekend. I've got the
> feeling that this is a task that requires a longer period of time.
> It's in SOMEDAY state in my Org along with:

I've found that you can pick up enough ELISP on a weekend to start
hacking small snipplets right away. Of course, you'll be very slow in
the beginning, and reverting to message-debugging all the time, but over
time you'll get more proficient. And you get to solve your problems,
which should lead to higher productivity in the long run.

Anyway, thanks for the links.

Cheers,
Viktor

Re: Wish: switch active time-stamps to inactive when CANCELED

[Memnon Anon]

Nick Dokos <nicholas.dokos@hp.com> writes:

>> But unfortunately, I know that little of ELISP, that I am not able
>> to implement it by myself :-(
> Time to learn some then - and there is no better way than scratching
> your own itch :-)

Two days later, no reply so far. I'll give it a shot.
Karl, if you are still fiddling to make it work, ignore this posting 
and keep trying. :)

>>> [...] accomplish this with org-after-todo-state-change-hook:
>>> Write a function that checks if org-state is "CANCELED"
>>> [...] looks for active timestamps and calls
>>> org-toggle-timestamp-type on them.

Everything there, so...

--8<---------------cut here---------------start------------->8---
(defun my-org-active-to-inactive-ts ()
  "Toggle type on active timestamp in current heading if state was changed to 'CNCL'.
  
This function is supposed to be called by 'org-after-todo-state-change-hook.
This function expects only one active timestamp to be in the headline."

  (when (string= org-state "CNCL") ; 1.
    (save-excursion ; 2.
      (save-restriction
	(widen)
	(outline-back-to-heading t) ; 3.
	(narrow-to-region (point)
			  (or (outline-next-heading) (point-max)))
	(when (re-search-backward (org-re-timestamp 'active) nil t) ; 4.
	  (org-toggle-timestamp-type))))))
    
(add-hook 'org-after-todo-state-change-hook 'my-org-active-to-inactive-ts)
--8<---------------cut here---------------end--------------->8---

1.: Only when org-state is CNCL, do the rest.
2.: We are jumping around and narrowing, so let's save the way things are,
    first.
3.: Move point to the beginning of the current headline, make sure we 
    only touch it by narrowing to it. If there is no next headline,
    point-max will do.
4.: regex search for (only) one active timestamp. When there is one, point
    will end up there and we toggle it. 
    When there is none, search returns nil, nothing happens.

Memnon

OT: Learning ELISP (was: Wish: switch active time-stamps to inactive when CANCELED)

[Karl Voit]

* Memnon Anon <gegendosenfleisch@googlemail.com> wrote:
>
> Two days later, no reply so far. I'll give it a shot.
> Karl, if you are still fiddling to make it work, ignore this posting 
> and keep trying. :)

Well, the hook in combination with the ARCHIVE tag did resolve my
current issue quite nicely. However, I follow ELISP lessons with
great pleasure independent of my current issues :-) But thanks to my
time budget, it's low priority for now. :-(

I did accomplish some advancement in the last weeks. I could fix
some if/when issues in my configuration and I do get the feeling
that meanwhile, I am able to judge the quality or the side-effects
of ELISP snippets I copy/copied from other people. Even this limited
knowledge gives me way more power over my Emacs than before.

The core of ELISP is not that huge at all. So far, I tend to think
that one key aspect of programming in ELISP is knowing, what
functions are out there, waiting to be used to accomplish a task.
And learning/finding all these handy functions is not trivial to me.
Especially in such a huge project like Org-mode.


Even more off-topic: I cleaned out some very old things and
summarized my Emacs config in (currently) two files.[1] And I am
thinking of moving it to Emacs initialization with babel[2] but so
far, I am not sure, if there are that many advantages over the
current situation.

  1. https://github.com/novoid/dot-emacs
  2. http://orgmode.org/worg/org-contrib/babel/intro.html#literate-programming
-- 
Karl Voit

Re: OT: Learning ELISP

[Jonathan Arkell]

I am a huge advocate of using org files and literate programming in your
config files.

A few reasons why:
  - Makes it easy to logically group sections of your init and
configuration
  - Agenda tags search on your initialization file?  Yes please!
Extremely useful for those "cross-cutting" bits.  I have tags like
"keybinding", "osx" and "linux", and working on others as appropriate.
  - Add TODOs to your init file.

Here is my example, but I stand on the shoulders of giants:

https://github.com/jonnay/emagicians-starter-kit



__________________________________


Jonathan Arkell
Sr. Developer
Inspired By Drum & Bass, Scheme, Kawaii

p. 403.206.4377
1011 9th Ave SE, Suite 300

Calgary, AB, Canada T2G 0Y4

jonathana@criticalmass.com
criticalmass.com





On 21/01/13 4:39 AM, "Karl Voit" <devnull@Karl-Voit.at> wrote:

>* Memnon Anon <gegendosenfleisch@googlemail.com> wrote:
>>
>> Two days later, no reply so far. I'll give it a shot.
>> Karl, if you are still fiddling to make it work, ignore this posting
>> and keep trying. :)
>
>Well, the hook in combination with the ARCHIVE tag did resolve my
>current issue quite nicely. However, I follow ELISP lessons with
>great pleasure independent of my current issues :-) But thanks to my
>time budget, it's low priority for now. :-(
>
>I did accomplish some advancement in the last weeks. I could fix
>some if/when issues in my configuration and I do get the feeling
>that meanwhile, I am able to judge the quality or the side-effects
>of ELISP snippets I copy/copied from other people. Even this limited
>knowledge gives me way more power over my Emacs than before.
>
>The core of ELISP is not that huge at all. So far, I tend to think
>that one key aspect of programming in ELISP is knowing, what
>functions are out there, waiting to be used to accomplish a task.
>And learning/finding all these handy functions is not trivial to me.
>Especially in such a huge project like Org-mode.
>
>
>Even more off-topic: I cleaned out some very old things and
>summarized my Emacs config in (currently) two files.[1] And I am
>thinking of moving it to Emacs initialization with babel[2] but so
>far, I am not sure, if there are that many advantages over the
>current situation.
>
>  1. https://github.com/novoid/dot-emacs
>  2.
>http://orgmode.org/worg/org-contrib/babel/intro.html#literate-programming
>--
>Karl Voit
>
>


________________________________

The information contained in this message is confidential. It is intended to be read only by the individual or entity named above or their designee. If the reader of this message is not the intended recipient, you are hereby notified that any distribution of this message, in any form, is strictly prohibited. If you have received this message in error, please immediately notify the sender and delete or destroy any copy of this message.

Using Org/babel for Emacs config files (was: OT: Learning ELISP)

[Karl Voit]

Hi!

* Jonathan Arkell <jonathana@criticalmass.com> wrote:
> I am a huge advocate of using org files and literate programming in your
> config files.

Great.

I am not completely convinced to convert my config to org/babel but
I am not oppose either. Not sure, if there are that many advantages,
that it is worth my effort, that's all.

> A few reasons why:
>   - Makes it easy to logically group sections of your init and
> configuration

Well, this is also possible with pure elisp and comments. Folding is
a different topic tough.

>   - Agenda tags search on your initialization file?  Yes please!
> Extremely useful for those "cross-cutting" bits.  I have tags like
> "keybinding", "osx" and "linux", and working on others as appropriate.

Hm. Interesting but I guess this is not my use-case (except for
:keybinding:). I do use "(when (system-is-linux) ... )" and similar
all over my configuration.

Although I am a huge fan of tagging, I don't see the use when
editing my emacs config (yet).

>   - Add TODOs to your init file.

I am using ~/org/mainfile.org -> "* shorts" -> "** TODO do this
:hostname:" all the time. Also for my config tasks. No worries: I do
see your point but I like to have all my sysadmin-tasks combined in
"mainfile.org/shorts". So maybe this is just me.

> Here is my example, but I stand on the shoulders of giants:
>
> https://github.com/jonnay/emagicians-starter-kit

Thanks for sharing!

One thing which I would appreciate would be that it is more easy to
write documentation (including URLs; outside of elisp sections) in
org.


Naive question: Are there any performance issues with not being able
to compile my config.org to config.elc as it is possible with
config.el files?

-- 
Karl Voit

Re: Using Org/babel for Emacs config files

[David Bjergaard]

Hi,

> Naive question: Are there any performance issues with not being able
> to compile my config.org to config.elc as it is possible with
> config.el files?
I used an init.org file in my initialization, and as long as you haven't
recently modified your init.org, emacs won't re-tangle the org file (ie
its as if you have a config.el file that's getting loaded)  Of course,
you can set emacs up to automatically recompile the config.el file when
its generated so, there are literally no performance issues because you
can still compile elisp code to load with the exception of when you edit
your configuration.

    Dave


Karl Voit <devnull@Karl-Voit.at> writes:

> Hi!
>
> * Jonathan Arkell <jonathana@criticalmass.com> wrote:
>> I am a huge advocate of using org files and literate programming in your
>> config files.
>
> Great.
>
> I am not completely convinced to convert my config to org/babel but
> I am not oppose either. Not sure, if there are that many advantages,
> that it is worth my effort, that's all.
>
>> A few reasons why:
>>   - Makes it easy to logically group sections of your init and
>> configuration
>
> Well, this is also possible with pure elisp and comments. Folding is
> a different topic tough.
>
>>   - Agenda tags search on your initialization file?  Yes please!
>> Extremely useful for those "cross-cutting" bits.  I have tags like
>> "keybinding", "osx" and "linux", and working on others as appropriate.
>
> Hm. Interesting but I guess this is not my use-case (except for
> :keybinding:). I do use "(when (system-is-linux) ... )" and similar
> all over my configuration.
>
> Although I am a huge fan of tagging, I don't see the use when
> editing my emacs config (yet).
>
>>   - Add TODOs to your init file.
>
> I am using ~/org/mainfile.org -> "* shorts" -> "** TODO do this
> :hostname:" all the time. Also for my config tasks. No worries: I do
> see your point but I like to have all my sysadmin-tasks combined in
> "mainfile.org/shorts". So maybe this is just me.
>
>> Here is my example, but I stand on the shoulders of giants:
>>
>> https://github.com/jonnay/emagicians-starter-kit
>
> Thanks for sharing!
>
> One thing which I would appreciate would be that it is more easy to
> write documentation (including URLs; outside of elisp sections) in
> org.
>
>
> Naive question: Are there any performance issues with not being able
> to compile my config.org to config.elc as it is possible with
> config.el files?

Re: Using Org/babel for Emacs config files (was: OT: Learning ELISP)

[Darlan Cavalcante Moreira]

I also have my Emacs configuration in org mode and it is really great.
Before moving to an org-mode initialization file I had multiple elisp files
and the main initialization file loaded the individual files. But having
everything in one place with the outline and search powers of org-modemuch
better. You can even put todos there if you want.

There are many benefits of using tags in your initialization. You can, for
instance, put a "Keybinds" tag in all headlines that change some
keybindings. You can put a "Flymake" tag in all headlines configuring
flymake for the different programming languages you use. If you need to
locate the flymake configuration specific for python, for instance, all
that is necessary is to do a tag search for the Flymake and Python tags.

Note that Emacs does not evaluate your initialization org-mode file
directly. It tangles all elisp blocks in that file to an elisp file and
then load that elisp file as usual. The tangling can be slow if you have
many blocks, but it only does that once after editing the org-mode
file. Therefore, you don't get a performance hit by having your
initialization in an org-mode file.

You can also take advantage of this tangle process. For instance, I have a
"Themes" headline where each child headline configures Emacs with a
different theme. I set the TANGLE property of all of these headlines to
"no" except one. Then Emacs will use that theme. You can easily
enable/disable big parts of your configuration like this and I find it
easier then commenting and uncommenting different parts of an elisp
file. For visual feedback, I like to add the COMMENT keyword in the
headlines where I have set the TANGLE property to no.

--
Darlan

At Wed, 23 Jan 2013 15:40:35 +0100,
Karl Voit wrote:
> 
> Hi!
> 
> * Jonathan Arkell <jonathana@criticalmass.com> wrote:
> > I am a huge advocate of using org files and literate programming in your
> > config files.
> 
> Great.
> 
> I am not completely convinced to convert my config to org/babel but
> I am not oppose either. Not sure, if there are that many advantages,
> that it is worth my effort, that's all.
> 
> > A few reasons why:
> >   - Makes it easy to logically group sections of your init and
> > configuration
> 
> Well, this is also possible with pure elisp and comments. Folding is
> a different topic tough.
> 
> >   - Agenda tags search on your initialization file?  Yes please!
> > Extremely useful for those "cross-cutting" bits.  I have tags like
> > "keybinding", "osx" and "linux", and working on others as appropriate.
> 
> Hm. Interesting but I guess this is not my use-case (except for
> :keybinding:). I do use "(when (system-is-linux) ... )" and similar
> all over my configuration.
> 
> Although I am a huge fan of tagging, I don't see the use when
> editing my emacs config (yet).
> 
> >   - Add TODOs to your init file.
> 
> I am using ~/org/mainfile.org -> "* shorts" -> "** TODO do this
> :hostname:" all the time. Also for my config tasks. No worries: I do
> see your point but I like to have all my sysadmin-tasks combined in
> "mainfile.org/shorts". So maybe this is just me.
> 
> > Here is my example, but I stand on the shoulders of giants:
> >
> > https://github.com/jonnay/emagicians-starter-kit
> 
> Thanks for sharing!
> 
> One thing which I would appreciate would be that it is more easy to
> write documentation (including URLs; outside of elisp sections) in
> org.
> 
> 
> Naive question: Are there any performance issues with not being able
> to compile my config.org to config.elc as it is possible with
> config.el files?
> 
> -- 
> Karl Voit
> 
> 

Re: Using Org/babel for Emacs config files (was: OT: Learning ELISP)

[Kyle Machulis]

Building on these idea, eric schulte's emacs 24 starter kit actually
uses tags to configure what should and shouldn't be tangled, to make
things easily configurable by others (not really a feature most of us
are looking for in configs, but interesting nonetheless). Check out
the starter-kit-load function in
https://raw.github.com/eschulte/emacs24-starter-kit/master/starter-kit.org

On Wed, Jan 23, 2013 at 2:06 PM, Darlan Cavalcante Moreira
<darcamo@gmail.com> wrote:
>
> I also have my Emacs configuration in org mode and it is really great.
> Before moving to an org-mode initialization file I had multiple elisp files
> and the main initialization file loaded the individual files. But having
> everything in one place with the outline and search powers of org-modemuch
> better. You can even put todos there if you want.
>
> There are many benefits of using tags in your initialization. You can, for
> instance, put a "Keybinds" tag in all headlines that change some
> keybindings. You can put a "Flymake" tag in all headlines configuring
> flymake for the different programming languages you use. If you need to
> locate the flymake configuration specific for python, for instance, all
> that is necessary is to do a tag search for the Flymake and Python tags.
>
> Note that Emacs does not evaluate your initialization org-mode file
> directly. It tangles all elisp blocks in that file to an elisp file and
> then load that elisp file as usual. The tangling can be slow if you have
> many blocks, but it only does that once after editing the org-mode
> file. Therefore, you don't get a performance hit by having your
> initialization in an org-mode file.
>
> You can also take advantage of this tangle process. For instance, I have a
> "Themes" headline where each child headline configures Emacs with a
> different theme. I set the TANGLE property of all of these headlines to
> "no" except one. Then Emacs will use that theme. You can easily
> enable/disable big parts of your configuration like this and I find it
> easier then commenting and uncommenting different parts of an elisp
> file. For visual feedback, I like to add the COMMENT keyword in the
> headlines where I have set the TANGLE property to no.
>
> --
> Darlan
>
> At Wed, 23 Jan 2013 15:40:35 +0100,
> Karl Voit wrote:
>>
>> Hi!
>>
>> * Jonathan Arkell <jonathana@criticalmass.com> wrote:
>> > I am a huge advocate of using org files and literate programming in your
>> > config files.
>>
>> Great.
>>
>> I am not completely convinced to convert my config to org/babel but
>> I am not oppose either. Not sure, if there are that many advantages,
>> that it is worth my effort, that's all.
>>
>> > A few reasons why:
>> >   - Makes it easy to logically group sections of your init and
>> > configuration
>>
>> Well, this is also possible with pure elisp and comments. Folding is
>> a different topic tough.
>>
>> >   - Agenda tags search on your initialization file?  Yes please!
>> > Extremely useful for those "cross-cutting" bits.  I have tags like
>> > "keybinding", "osx" and "linux", and working on others as appropriate.
>>
>> Hm. Interesting but I guess this is not my use-case (except for
>> :keybinding:). I do use "(when (system-is-linux) ... )" and similar
>> all over my configuration.
>>
>> Although I am a huge fan of tagging, I don't see the use when
>> editing my emacs config (yet).
>>
>> >   - Add TODOs to your init file.
>>
>> I am using ~/org/mainfile.org -> "* shorts" -> "** TODO do this
>> :hostname:" all the time. Also for my config tasks. No worries: I do
>> see your point but I like to have all my sysadmin-tasks combined in
>> "mainfile.org/shorts". So maybe this is just me.
>>
>> > Here is my example, but I stand on the shoulders of giants:
>> >
>> > https://github.com/jonnay/emagicians-starter-kit
>>
>> Thanks for sharing!
>>
>> One thing which I would appreciate would be that it is more easy to
>> write documentation (including URLs; outside of elisp sections) in
>> org.
>>
>>
>> Naive question: Are there any performance issues with not being able
>> to compile my config.org to config.elc as it is possible with
>> config.el files?
>>
>> --
>> Karl Voit
>>
>>
>

[OT] open url in emacs was: Re: Using Org/babel for Emacs config files

[Andreas Leha]

Kyle Machulis <kyle@nonpolynomial.com> writes:

> Building on these idea, eric schulte's emacs 24 starter kit actually
> uses tags to configure what should and shouldn't be tangled, to make
> things easily configurable by others (not really a feature most of us
> are looking for in configs, but interesting nonetheless). Check out
> the starter-kit-load function in
> https://raw.github.com/eschulte/emacs24-starter-kit/master/starter-kit.org
>

Apologies for hijacking this thread.  But could someone tell me how to
open that link directly in emacs?  My try with 'browse-url-emacs'
failed.

Thanks in advance,
Andreas

Re: [OT] open url in emacs was: Re: Using Org/babel for Emacs config files

[Bastien]

Hi Andreas,

Andreas Leha <andreas.leha@med.uni-goettingen.de> writes:

> Apologies for hijacking this thread.  But could someone tell me how to
> open that link directly in emacs?  My try with 'browse-url-emacs'
> failed.

You cannot.  Get the file and open it.

Best,

-- 
 Bastien

Re: [OT] open url in emacs was: Re: Using Org/babel for Emacs config files

[Andreas Leha]

Hi Bastien,

Bastien <bzg@altern.org> writes:

> Hi Andreas,
>
> Andreas Leha <andreas.leha@med.uni-goettingen.de> writes:
>
>> Apologies for hijacking this thread.  But could someone tell me how to
>> open that link directly in emacs?  My try with 'browse-url-emacs'
>> failed.

Thanks for answering that OT question, although ...

> You cannot.  Get the file and open it.
  ^^^^^^^^^^  that is something I'd not expect to hear when it comes to
emacs....


- Andreas

Re: [OT] open url in emacs was: Re: Using Org/babel for Emacs config files

[Bastien]

Andreas Leha <andreas.leha@med.uni-goettingen.de> writes:

>   ^^^^^^^^^^  that is something I'd not expect to hear when it comes to
> emacs....

(when (switch-to-buffer-other-window
       (url-retrieve-synchronously
	"https://raw.github.com/eschulte/emacs24-starter-kit/master/starter-kit.org"))
  (write-file (read-from-minibuffer "File: "))
  (org-mode))

-- 
 Bastien

Re: [OT] open url in emacs was: Re: Using Org/babel for Emacs config files

[Andreas Leha]

Hi Bastien,

Bastien <bzg@altern.org> writes:

> Andreas Leha <andreas.leha@med.uni-goettingen.de> writes:
>
>>   ^^^^^^^^^^  that is something I'd not expect to hear when it comes to
>> emacs....
>
> (when (switch-to-buffer-other-window
>        (url-retrieve-synchronously
> 	"https://raw.github.com/eschulte/emacs24-starter-kit/master/starter-kit.org"))
>   (write-file (read-from-minibuffer "File: "))
>   (org-mode))

;-)

Thanks a lot for that!  Works nicely

Regards,
Andreas

Re: [OT] open url in emacs was: Re: Using Org/babel for Emacs config files

[Jambunathan K]

Bastien <bzg@altern.org> writes:

> Andreas Leha <andreas.leha@med.uni-goettingen.de> writes:
>
>>   ^^^^^^^^^^  that is something I'd not expect to hear when it comes to
>> emacs....
>
> (when (switch-to-buffer-other-window
>        (url-retrieve-synchronously
> 	"https://raw.github.com/eschulte/emacs24-starter-kit/master/starter-kit.org"))
>   (write-file (read-from-minibuffer "File: "))
>   (org-mode))

M-x browse-url-emacs RET

-- 

Re: [OT] open url in emacs was: Re: Using Org/babel for Emacs config files

[Andreas Leha]

Hi Jambu,

Jambunathan K <kjambunathan@gmail.com> writes:

> Bastien <bzg@altern.org> writes:
>
>> Andreas Leha <andreas.leha@med.uni-goettingen.de> writes:
>>
>>>   ^^^^^^^^^^  that is something I'd not expect to hear when it comes to
>>> emacs....
>>
>> (when (switch-to-buffer-other-window
>>        (url-retrieve-synchronously
>> 	"https://raw.github.com/eschulte/emacs24-starter-kit/master/starter-kit.org"))
>>   (write-file (read-from-minibuffer "File: "))
>>   (org-mode))
>
> M-x browse-url-emacs RET

thanks, I tried that, but that does not seem to work for me:

(sorry for the German messaeges)
,----
| Contacting host: raw.github.com:443 [3 times]
| Note: file is write protected
| Contacting host: raw.github.com:443
| Loading reftex...done
| Contacting host: raw.github.com:443
| Compiling label environment definitions...done
| Contacting host: raw.github.com:443
| Scanning entire document...
| No such file: https://raw.github.com/eschulte/emacs24-starter-kit/master/starter-kit.org (ignored)
| Scanning file https://raw.github.com/eschulte/emacs24-starter-kit/master/starter-kit.org
| Scanning document... done
| Local Ispell dictionary set to english
| Contacting host: raw.github.com:443 [3 times]
| open-network-stream: /443 Zu diesem Hostnamen gehört keine Adresse
`----

if you tell me how to make it work, that would be even simpler.

Cheers,
Andreas

Re: [OT] open url in emacs was: Re: Using Org/babel for Emacs config files

[Jambunathan K]

Andreas Leha <andreas.leha@med.uni-goettingen.de> writes:

> Hi Jambu,
>
> Jambunathan K <kjambunathan@gmail.com> writes:
>
>> Bastien <bzg@altern.org> writes:
>>
>>> Andreas Leha <andreas.leha@med.uni-goettingen.de> writes:
>>>
>>>>   ^^^^^^^^^^  that is something I'd not expect to hear when it comes to
>>>> emacs....
>>>
>>> (when (switch-to-buffer-other-window
>>>        (url-retrieve-synchronously
>>> 	"https://raw.github.com/eschulte/emacs24-starter-kit/master/starter-kit.org"))
>>>   (write-file (read-from-minibuffer "File: "))
>>>   (org-mode))
>>
>> M-x browse-url-emacs RET

It works for me here without any extra props.

I would assume it has something to do with 's' in https.  It is possible
that that your local Emacs is not able to do TLS.  Change `https://' to
`http://'.

I run Emacs from trunk.  May be your Emacs is old - for some values of
old.

> thanks, I tried that, but that does not seem to work for me:
>
> (sorry for the German messaeges)
> ,----
> | Contacting host: raw.github.com:443 [3 times]
> | Note: file is write protected
> | Contacting host: raw.github.com:443
> | Loading reftex...done
> | Contacting host: raw.github.com:443
> | Compiling label environment definitions...done
> | Contacting host: raw.github.com:443
> | Scanning entire document...
> | No such file:
> | https://raw.github.com/eschulte/emacs24-starter-kit/master/starter-kit.org
> | (ignored)
> | Scanning file
> | https://raw.github.com/eschulte/emacs24-starter-kit/master/starter-kit.org
> | Scanning document... done
> | Local Ispell dictionary set to english
> | Contacting host: raw.github.com:443 [3 times]
> | open-network-stream: /443 Zu diesem Hostnamen gehört keine Adresse
> `----

> if you tell me how to make it work, that would be even simpler.
>
> Cheers,
> Andreas
>
>
>

-- 

Re: [OT] open url in emacs was: Re: Using Org/babel for Emacs config files

[Andreas Leha]

Jambunathan K <kjambunathan@gmail.com> writes:

> Andreas Leha <andreas.leha@med.uni-goettingen.de> writes:
>
>> Hi Jambu,
>>
>> Jambunathan K <kjambunathan@gmail.com> writes:
>>
>>> Bastien <bzg@altern.org> writes:
>>>
>>>> Andreas Leha <andreas.leha@med.uni-goettingen.de> writes:
>>>>
>>>>>   ^^^^^^^^^^  that is something I'd not expect to hear when it comes to
>>>>> emacs....
>>>>
>>>> (when (switch-to-buffer-other-window
>>>>        (url-retrieve-synchronously
>>>> 	"https://raw.github.com/eschulte/emacs24-starter-kit/master/starter-kit.org"))
>>>>   (write-file (read-from-minibuffer "File: "))
>>>>   (org-mode))
>>>
>>> M-x browse-url-emacs RET
>
> It works for me here without any extra props.
>
> I would assume it has something to do with 's' in https.  It is possible
> that that your local Emacs is not able to do TLS.  Change `https://' to
> `http://'.
>
> I run Emacs from trunk.  May be your Emacs is old - for some values of
> old.
>

I finally tracked this down.

Disabling egg (from https://github.com/byplayer/egg.git) solved my
problem with browse-url-emacs.
There seems to be a(nother) problem in egg...

Thank to all looking into this.

Regards,
Andreas

Re: [OT] open url in emacs was: Re: Using Org/babel for Emacs config files

[Daimrod]

[-- Attachment #1: Type: text/plain, Size: 1060 bytes --]

Andreas Leha <andreas.leha@med.uni-goettingen.de> writes:

> Kyle Machulis <kyle@nonpolynomial.com> writes:
>> https://raw.github.com/eschulte/emacs24-starter-kit/master/starter-kit.org
> Apologies for hijacking this thread.  But could someone tell me how to
> open that link directly in emacs?  My try with 'browse-url-emacs'
> failed.

You can use Emacs-w3m[1] to browse the web from Emacs.

BTW, I've made functions to browse and copy URL in Emacs[2][3].

- `C-c b' to browse the URL at point
- `C-c y' to copy the URL at point

You can customize which browser is used to browse an URL with
`gu-first-browse-url-function' and `gu-second-browse-url-function' (by
default it uses `browse-url-browser-function' and `browse-url-generic').
The first one is used by default, the second one when you use a prefix
(`C-u C-c b').

[1] http://emacs-w3m.namazu.org/
[2] https://github.com/daimrod/Emacs-config/blob/master/elisp/graze-url.el
[3] https://github.com/daimrod/Emacs-config/blob/master/config-bindings.el#L73

-- 
Daimrod/Greg

[-- Attachment #2: Type: application/pgp-signature, Size: 835 bytes --]

Re: Using Org/babel for Emacs config files

[Thorsten Jolitz]

Karl Voit <devnull@Karl-Voit.at> writes:

Hi, 

> I am not completely convinced to convert my config to org/babel but
> I am not oppose either. Not sure, if there are that many advantages,
> that it is worth my effort, that's all.

its a bit like writing with LaTeX - in some simpler but common cases its
just great to write your stuff in Org and simply press a few keys for
the LaTeX export. But dealing with LaTeX AND Org-mode when writing a big
complex document might not be such a good idea, there will be enough
problems with LaTeX only. 

I must admit I went back from Org-based multi-file config to a single
big .el init file because of two reasons:

- I frequently found myself looking for stuff in the tangled .el files
  instead of the .org files, because I was only interested in the lisp
  code and did not want to see all those comments in the .org files. 
- after figuring out that I can have full outline functionality in .el
  files too, the (in my eyes) main advantage of an org-based config was
  gone. 

Otherwise I really liked the org based starter-kit, it was well
organized and definitely better than what I had before. 

-- 
cheers,
Thorsten

Re: Using Org/babel for Emacs config files

[Thomas S. Dye]

Thorsten Jolitz <tjolitz@googlemail.com> writes:

> Karl Voit <devnull@Karl-Voit.at> writes:
>
> Hi, 
>
>> I am not completely convinced to convert my config to org/babel but
>> I am not oppose either. Not sure, if there are that many advantages,
>> that it is worth my effort, that's all.
>
> its a bit like writing with LaTeX - in some simpler but common cases its
> just great to write your stuff in Org and simply press a few keys for
> the LaTeX export. But dealing with LaTeX AND Org-mode when writing a big
> complex document might not be such a good idea, there will be enough
> problems with LaTeX only. 

I wrote a book and several articles in Org mode using the old exporter.
Interesting exercises, but you're right that it would have been easier
in straight LaTeX (except for the babel parts).

I've been working with Nicolas Goaziou's new LaTeX exporter recently and
the gap between Org mode and AucTeX has narrowed considerably.
Recently, I wrote a paper for a journal that accepts LaTeX, but doesn't
have a LaTeX class file, only a template that uses the standard article
class and prescribes which packages must be used, with no additions.  It
turned out to be easy to configure the new exporter, even in this
difficult circumstance.  

The new exporter is *really* nice.  It feels right, even for big complex
documents.  I'm still using LaTeX at work, but I now write exclusively
in Org mode for all my projects outside work.  The balance has tipped
for me.
  
All the best,
Tom

-- 
Thomas S. Dye
http://www.tsdye.com

Re: Using Org/babel for Emacs config files

[Thorsten Jolitz]

tsd@tsdye.com (Thomas S. Dye) writes:

> I've been working with Nicolas Goaziou's new LaTeX exporter recently and
> the gap between Org mode and AucTeX has narrowed considerably.

Thats of course always true for Org-mode, and often the problem isn't
really in the software, but sits in front of the computer and is not
up-to-date or forgot how to do it. 

> The new exporter is *really* nice. 

Agreed ... 

-- 
cheers,
Thorsten

full outline functionality in .el files: how? (was: Re: Using Org/babel for Emacs config files)

[Gregor Zattler]

Hi Thorsten,
* Thorsten Jolitz <tjolitz@googlemail.com> [24. Jan. 2013]:
> - after figuring out that I can have full outline functionality in .el
>   files too, the (in my eyes) main advantage of an org-based config was
>   gone. 

could you please elaborate how to achieve this?

Ciao, Gregor
-- 
 -... --- .-. . -.. ..--.. ...-.-

Re: full outline functionality in .el files: how?

[Bastien]

Hi Gregor,

Gregor Zattler <telegraph@gmx.net> writes:

> Hi Thorsten,
> * Thorsten Jolitz <tjolitz@googlemail.com> [24. Jan. 2013]:
>> - after figuring out that I can have full outline functionality in .el
>>   files too, the (in my eyes) main advantage of an org-based config was
>>   gone. 
>
> could you please elaborate how to achieve this?

See here: http://orgmode.org/worg/org-faq.html#sec-5-4

-- 
 Bastien

Re: full outline functionality in .el files: how?

[Thorsten Jolitz]

Bastien <bzg@altern.org> writes:

Hi Bastien,

> Gregor Zattler <telegraph@gmx.net> writes:
>
>> Hi Thorsten,
>> * Thorsten Jolitz <tjolitz@googlemail.com> [24. Jan. 2013]:
>>> - after figuring out that I can have full outline functionality in .el
>>>   files too, the (in my eyes) main advantage of an org-based config was
>>>   gone. 
>>
>> could you please elaborate how to achieve this?
>
> See here: http://orgmode.org/worg/org-faq.html#sec-5-4

I just tried it, the org-cyle commands do work in my init.el (in
outline-minor-mode), except the global cycling (with one C-u), that
one does nothing. But the others (zero Args, or with C-u C-u or C-u C-u
C-u) seem to work as expected. 

-- 
cheers,
Thorsten

Re: full outline functionality in .el files: how?

[Bastien]

Hi Thorsten,

Thorsten Jolitz <tjolitz@googlemail.com> writes:

> I just tried it, the org-cyle commands do work in my init.el (in
> outline-minor-mode), except the global cycling (with one C-u), that
> one does nothing. But the others (zero Args, or with C-u C-u or C-u C-u
> C-u) seem to work as expected. 

I don't understand what does not work exactly -- can you tell me?

-- 
 Bastien

Re: full outline functionality in .el files: how?

[Thorsten Jolitz]

Bastien <bzg@altern.org> writes:

Hi Bastien,

> I don't understand what does not work exactly -- can you tell me?

ok, I checked more exactly in my init.el in outline-minor-mode:

(1)

M-x org-cycle on the top-level entries expands the entry one level, and
then collapses it again, i.e. cycling between two states only

(2)

C-u M-x org-cyle does nothing

(3)

C-u C-u M-x org-cyle actually does the global cycling through 3 states,
but it goes from OVERVIEW to SHOW ALL to CONTENTS (1. to 3. to 2.)

(4)

C-u C-u C-u M-x org-cyle makes the entire buffer visible the first time,
and does nothing when invoked again. 
 
-- 
cheers,
Thorsten

Re: full outline functionality in .el files: how?

[Bastien]

Oh!  Now I understand why we miscommunicate...

You checked this link...


  http://orgmode.org/worg/org-faq.html#sec-5-4

But the webpage was perhaps not up to date and showed this
link instead:

  http://orgmode.org/worg/org-faq.html#sec-5-5

See the first link for my answer to Gregor, which
is about using orgstruct-mode folding capacity in
emacs-lisp-mode.

Also, the difference between Org's cycling and
those local cycling features may be related to the
fact that the org-cycle functions are not called
interactively -- at least in the first FAQ above.

Best,

-- 
 Bastien

Re: full outline functionality in .el files: how?

[Thorsten Jolitz]

Gregor Zattler <telegraph@gmx.net> writes:

Hi Gregor,

> * Thorsten Jolitz <tjolitz@googlemail.com> [24. Jan. 2013]:
>> - after figuring out that I can have full outline functionality in .el
>>   files too, the (in my eyes) main advantage of an org-based config was
>>   gone. 
>
> could you please elaborate how to achieve this?

I already shared my new trick (which might not be that new for many
others) on the PicoLisp mailing list:

,-------------------------------------------------------------------
| http://www.mail-archive.com/picolisp@software-lab.de/msg03764.html
`-------------------------------------------------------------------

When you replace # with ; as comment char, there is no difference
between a PicoLisp source file (.l) or an Emacs Lisp source file (.el)
with regards to this trick. 

Actually, the Org-mode visibility commands mostly work out of the box
too, without the .l or .el source buffer being in org (minor) mode,
since outline-minor-mode is active and these Org commands often simply
act on the (buffer local) 'outline-regexp', which might be '^;; \*+ ' or
'^## \*+ ' or something similar too.

-- 
cheers,
Thorsten

Re: full outline functionality in .el files: how?

[Memnon Anon]

Gregor Zattler <telegraph@gmx.net> writes:

> Hi Thorsten,
>> - after figuring out that I can have full outline functionality in .el
>>   files too, the (in my eyes) main advantage of an org-based config was
>>   gone. 
>
> could you please elaborate how to achieve this?

Icicle uses linkd.el:
http://www.emacswiki.org/cgi-bin/wiki/linkd.el

Memnon