[contacts] No completion in Gnus

[contacts] No completion in Gnus

[Thorsten]

Hello List, 
I have a org-contacts file with a few contacts with EMAIL property,
taken from gnus using org-capture as described in the org-contacts
manual. The contacts are tagged with some tags. But if I try to send an
email to one of the contacts using completion in message-mode, nothing
happens. I type + and press <TAB> in the To: header, but to no avail. I
even restarted Emacs without loading bbdb, since there seemed to be some
interference, but that didn't help either. I cannot find any more variables to
connect org-contacts with gnus, so I guess it should work out of the
box. 

What might be wrong in this case?
Thanks for any hints. 

PS
One further org-contacts related question: there is a predefined
property ADDRESS, without any inner structure. Am I supposed to write my own
org-capture template for that property? How would that look like - just
one single string? If I want something more structured (street, city,
zip-code etc), I have to define one property for each info-piece, or is
there something like a compound property? 

-- 
Thorsten

Re: [contacts] No completion in Gnus

[Jambunathan K]

Thorsten <quintfall@googlemail.com> writes:

> Hello List, 
> I have a org-contacts file with a few contacts with EMAIL property,
> taken from gnus using org-capture as described in the org-contacts
> manual. The contacts are tagged with some tags. But if I try to send an
> email to one of the contacts using completion in message-mode, nothing
> happens. I type + and press <TAB> in the To: header, but to no avail. I
> even restarted Emacs without loading bbdb, since there seemed to be some
> interference, but that didn't help either. I cannot find any more variables to
> connect org-contacts with gnus, so I guess it should work out of the
> box. 

(caveat: I have never used org-contacts)

In such cases, typically there is an insinuate function. Yes, there is
one

(org-contacts-gnus-insinuate)

Try adding that to .emacs or eval it.

I also see this snippet of code in org-contacts.el and is done
automatically for you.

(when (boundp 'completion-at-point-functions)
  (add-hook 'message-mode-hook
	    (lambda ()
	      (add-to-list 'completion-at-point-functions
			   'org-contacts-message-complete-function))))


So while composing the mail, check the value of variable
completion-at-point-functions and make sure you see some org-contacts
related entries there.

Looks like the guts of completion happen in
org-contacts-complete-name. So you can place your cursor on a message
header and try doing

M-: (org-contacts-complete-name)

Hope something interesting will happen.

> What might be wrong in this case?
> Thanks for any hints. 
>
> PS
> One further org-contacts related question: there is a predefined
> property ADDRESS, without any inner structure. Am I supposed to write my own
> org-capture template for that property? How would that look like - just
> one single string? If I want something more structured (street, city,
> zip-code etc), I have to define one property for each info-piece, or is
> there something like a compound property? 

-- 

Re: [contacts] No completion in Gnus

[Thorsten]

Jambunathan K <kjambunathan@gmail.com> writes:

> Thorsten <quintfall@googlemail.com> writes:
>
>> Hello List, I have a org-contacts file with a few contacts with EMAIL
>> property, taken from gnus using org-capture as described in the
>> org-contacts manual. The contacts are tagged with some tags. But if I
>> try to send an email to one of the contacts using completion in
>> message-mode, nothing happens. I type + and press <TAB> in the To:
>> header, but to no avail. I even restarted Emacs without loading bbdb,
>> since there seemed to be some interference, but that didn't help
>> either. I cannot find any more variables to connect org-contacts with
>> gnus, so I guess it should work out of the box.
>
> (caveat: I have never used org-contacts)
>
> In such cases, typically there is an insinuate function. Yes, there is
> one
>
> (org-contacts-gnus-insinuate)
>
> Try adding that to .emacs or eval it.
>
> I also see this snippet of code in org-contacts.el and is done
> automatically for you.
>
> (when (boundp 'completion-at-point-functions)
>   (add-hook 'message-mode-hook
> 	    (lambda ()
> 	      (add-to-list 'completion-at-point-functions
> 			   'org-contacts-message-complete-function))))
>
>
> So while composing the mail, check the value of variable
> completion-at-point-functions and make sure you see some org-contacts
> related entries there.

Thats the culprit - I use Emacs 23.1.1, and it seems that variable
completion-at-point-functions has been added to Emacs just recently
with version 23.2. So I have to update my Emacs - should I go for Emacs
24 right away, or is that to early?


> Looks like the guts of completion happen in
> org-contacts-complete-name. So you can place your cursor on a message
> header and try doing
>
> M-: (org-contacts-complete-name)
>
> Hope something interesting will happen.
>
>> What might be wrong in this case?
>> Thanks for any hints. 
>>
>> PS
>> One further org-contacts related question: there is a predefined
>> property ADDRESS, without any inner structure. Am I supposed to write my own
>> org-capture template for that property? How would that look like - just
>> one single string? If I want something more structured (street, city,
>> zip-code etc), I have to define one property for each info-piece, or is
>> there something like a compound property? 


Thanks for your help

-- 
Thorsten

Re: [contacts] No completion in Gnus

[Achim Gratz]

Thorsten <quintfall@googlemail.com> writes:
> Thats the culprit - I use Emacs 23.1.1, and it seems that variable
> completion-at-point-functions has been added to Emacs just recently
> with version 23.2. So I have to update my Emacs - should I go for Emacs
> 24 right away, or is that to early?

The current release version for Emacs is 23.3.

I've switched to Emacs24 about half a year ago (they are just now going
into pre-test, so things should stabilize) and it has been smooth
sailing for me, but I'd recommend to keep the 23.3 version around as a
fallback.  There are quite a few changes vs. Emacs23 that you may have
to think about, read the NEWS file.  I've compiled Emacs24 myself (from
Git), most pre-compiled packages represent a much older state from the
beginning of this year.  I've backed out a dodgy change once or twice
and this is so much easier to do in Git; even though there would have
been other workarounds.


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

Waldorf MIDI Implementation & additional documentation:
http://Synth.Stromeko.net/Downloads.html#WaldorfDocs

Re: [contacts] No completion in Gnus

[Jason Dunsmore]

I'm trying to get org-contacts completion working using Jambunathan's
instructions...

Jambunathan K <kjambunathan@gmail.com> writes:

> In such cases, typically there is an insinuate function. Yes, there is
> one
>
> (org-contacts-gnus-insinuate)

I've added this function to my config and I'm still not seeing email
address completion in Gnus.

When I do an "M-x org-contacts" search for one of my contacts, I get the
expected results.

> So while composing the mail, check the value of variable
> completion-at-point-functions and make sure you see some org-contacts
> related entries there.

My completion-at-point-functions's value is
(org-contacts-message-complete-function)

> Looks like the guts of completion happen in
> org-contacts-complete-name. So you can place your cursor on a message
> header and try doing
>
> M-: (org-contacts-complete-name)

When I do this, I see:

(5 8 (lambda (&rest --cl-rest--) (apply #[... "ÃÄJ
\"\"‡" [G73503 G73502 args2 apply append] 5] (quote --args1--) (quote --fun--) --cl-rest--)))

Is there anything else I'm missing?

Thanks,
Jason

Re: [contacts] No completion in Gnus

[Julien Danjou]

[-- Attachment #1: Type: text/plain, Size: 134 bytes --]

On Wed, Oct 05 2011, Jason Dunsmore wrote:

> Is there anything else I'm missing?

Are you using Emacs 24?

-- 
Julien Danjou

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

Re: [contacts] No completion in Gnus

[jasondunsmore]

Julien Danjou <julien@danjou.info> writes:

> On Wed, Oct 05 2011, Jason Dunsmore wrote:
>
>> Is there anything else I'm missing?
>
> Are you using Emacs 24?

I'm running Emacs 23.2.1.  Does it require Emacs 24?

Re: [contacts] No completion in Gnus

[Julien Danjou]

[-- Attachment #1: Type: text/plain, Size: 141 bytes --]

On Wed, Oct 05 2011, jasondunsmore@gmail.com wrote:

> I'm running Emacs 23.2.1.  Does it require Emacs 24?

Yes.

-- 
Julien Danjou

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

Re: [contacts] No completion in Gnus

[Thorsten]

Achim Gratz <Stromeko@nexgo.de> writes:

> The current release version for Emacs is 23.3.
>
> I've switched to Emacs24 about half a year ago (they are just now going
> into pre-test, so things should stabilize) and it has been smooth
> sailing for me, but I'd recommend to keep the 23.3 version around as a
> fallback.  There are quite a few changes vs. Emacs23 that you may have
> to think about, read the NEWS file.  I've compiled Emacs24 myself (from
> Git), most pre-compiled packages represent a much older state from the
> beginning of this year.  I've backed out a dodgy change once or twice
> and this is so much easier to do in Git; even though there would have
> been other workarounds.

Since I always expect unexpected things to happen when upgrading, I
would rather go directly to Emacs 24 when its more or less stable
already - and thats what it seems to be.

I might use the opportunity to change from 32bit to 64bit, from Ubuntu
to Arch Linux (which looks good) and from Gnome to Stump-WM. That looks
like a nice combination to me. 

Cheers
-- 
Thorsten

Re: [contacts] No completion in Gnus

[Michael Markert]

[-- Attachment #1: Type: text/plain, Size: 783 bytes --]

On  5 Oct 2011, jasondunsmore@gmail.com wrote:

> Julien Danjou <julien@danjou.info> writes:
>
>> On Wed, Oct 05 2011, Jason Dunsmore wrote:
>>
>>> Is there anything else I'm missing?
>>
>> Are you using Emacs 24?
>
> I'm running Emacs 23.2.1.  Does it require Emacs 24?

As Julien said: It does require Emacs 24.

But you can bind `org-contacts-message-complete-function' directly to a
key (You need to wrap it though) and bypass
`completion-at-point-functions':

#+begin_src emacs-lisp

(add-to-hook 'message-mode-hook
             (lambda () (local-set-key (kbd "TAB")
                                       (lambda ()
                                         (interactive)
                                         (org-contacts-message-complete-function)))))

#+end_src

Michael

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

org-contacts: how to manage contacts; structure of meta data (was: No completion in Gnus)

[Karl Voit]

* Thorsten <quintfall@googlemail.com> wrote:
>
> I have a org-contacts file with a few contacts with EMAIL property,
> taken from gnus using org-capture as described in the org-contacts
> manual. The contacts are tagged with some tags. 

[...]

> PS
> One further org-contacts related question: there is a predefined
> property ADDRESS, without any inner structure. Am I supposed to write my own
> org-capture template for that property? How would that look like - just
> one single string? If I want something more structured (street, city,
> zip-code etc), I have to define one property for each info-piece, or is
> there something like a compound property? 

I moved all of my approx. 1000 contacts from jPilot contacts to
org-contacts a few months ago. I faced the same questions you are
mentioning since org-contacts[1] does only mention the :EMAIL:
property.

A short research with my favourite search engine did not result in
any (ISO?) standard that relates to «how to define/describe contact
meta data».

Without describing further attempts, I ended up with following
yasnippet[2] template for a new contact:

,----
| # name : Org-contacts template for a person or a company
| # --
| ** $1 $2  :$1$2:
| :PROPERTIES:
| :TYPE: ${3:$$(yas/choose-value '("person" "company"))}
| :TITLE:
| :EMAIL: $4
| :URL:
| :MOBILE: 0043/
| :HOMEPHONE:
| :WORKPHONE:
| :PHONE:
| :COMPANY:
| :STREET:
| :POSTALCODE:
| :CITY:
| :COUNTRY: Österreich
| :END:
|
| first contact: $0
`----

My current tags for contacts are defined in the header of my
contacts.org:

,----[ first line of my contacts.org ]
| #+TAGS: job(j) friends(f) health(e) hotels(o) restaurants_bars(r) \
|         sport(s) students_TU(t) relatives(r) company(c)
`----

... but this is pretty much related to my previous system
(jPilot/datebk6/PalmOS) which offered only one single category for
each contact. I might re-think these tags in future.


A very handy bonus feature is referencing my contacts. In my .emacs
I do have following:

,----[ .emacs (excerpt) ]
| (setq org-link-abbrev-alist
|       '(
|     ("contact" . "~/org/contacts.org::/\*.*%s/")
|     ))
`----

Combined with following yasnippet template it results in very comfty
contact handling:

,----[ contact yasnippet ]
| # name : expand link to contact
| # --
| [[contact:$1][${2:$$(unless yas/modified-p
|  (let ((field (nth 0 (yas/snippet-fields (first (yas/snippets-at-point))))))
|    (concat (buffer-substring (yas/field-start field) (yas/field-end field)))))}]] $0
`----

Basically, you have to enter the name of the snippet, press «TAB»
and then you type in the name (or anything of the heading line of a
contact). Then it results in [[contact:foo bar][foo bar]] which is a
clickable link to any contact which relates to «foo bar».

Whenever I rename my contacts.org or put it in another folder, I
just have to modify the line in my .emacs and not each link.

I like this system.

HTH

  1. http://julien.danjou.info/org-contacts.html
  2. http://code.google.com/p/yasnippet/
-- 
Karl Voit

Re: org-contacts: how to manage contacts; structure of meta data (was: No completion in Gnus)

[Thorsten]

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

>> One further org-contacts related question: there is a predefined
>> property ADDRESS, without any inner structure. Am I supposed to write my own
>> org-capture template for that property? How would that look like - just
>> one single string? If I want something more structured (street, city,
>> zip-code etc), I have to define one property for each info-piece, or is
>> there something like a compound property? 
>
> I moved all of my approx. 1000 contacts from jPilot contacts to
> org-contacts a few months ago. I faced the same questions you are
> mentioning since org-contacts[1] does only mention the :EMAIL:
> property.
>
> A short research with my favourite search engine did not result in
> any (ISO?) standard that relates to «how to define/describe contact
> meta data».
>
> Without describing further attempts, I ended up with following
> yasnippet[2] template for a new contact:

Do you think that yasnippet is superior to org-capture or is it just a
casuality that you prefered it to org-capture?


> ,----
> | # name : Org-contacts template for a person or a company
> | # --
> | ** $1 $2  :$1$2:
> | :PROPERTIES:
> | :TYPE: ${3:$$(yas/choose-value '("person" "company"))}
> | :TITLE:
> | :EMAIL: $4
> | :URL:
> | :MOBILE: 0043/
> | :HOMEPHONE:
> | :WORKPHONE:
> | :PHONE:
> | :COMPANY:
> | :STREET:
> | :POSTALCODE:
> | :CITY:
> | :COUNTRY: Österreich
> | :END:
> |
> | first contact: $0
> `----

Thats what I thought, that a single :ADDRESS: property is not very
usefull or practical. So there is not something like a nested property,
each piece of information needs its own property. I think your list is
pretty complete and usable. 


> My current tags for contacts are defined in the header of my
> contacts.org:
>
> ,----[ first line of my contacts.org ]
> | #+TAGS: job(j) friends(f) health(e) hotels(o) restaurants_bars(r) \
> |         sport(s) students_TU(t) relatives(r) company(c)
> `----

[how do you produce this nice insertions in your email?]

> A very handy bonus feature is referencing my contacts. In my .emacs
> I do have following:
>
> ,----[ .emacs (excerpt) ]
> | (setq org-link-abbrev-alist
> |       '(
> |     ("contact" . "~/org/contacts.org::/\*.*%s/")
> |     ))
> `----

> Combined with following yasnippet template it results in very comfty
> contact handling:
>
> ,----[ contact yasnippet ]
> | # name : expand link to contact
> | # --
> | [[contact:$1][${2:$$(unless yas/modified-p
> |  (let ((field (nth 0 (yas/snippet-fields (first (yas/snippets-at-point))))))
> |    (concat (buffer-substring (yas/field-start field) (yas/field-end field)))))}]] $0
> `----
>
> Basically, you have to enter the name of the snippet, press «TAB»
> and then you type in the name (or anything of the heading line of a
> contact). Then it results in [[contact:foo bar][foo bar]] which is a
> clickable link to any contact which relates to «foo bar».

Great, I'll give that a try too.

Thanks for your help
Cheers

-- 
Thorsten

Re: org-contacts: how to manage contacts; structure of meta data

[Karl Voit]

* Thorsten <quintfall@googlemail.com> wrote:
> Karl Voit <devnull@Karl-Voit.at> writes:
>
>> Without describing further attempts, I ended up with following
>> yasnippet[2] template for a new contact:
>
> Do you think that yasnippet is superior to org-capture or is it just a
> casuality that you prefered it to org-capture?

Oh, first I started with org-capture. One week later I started to
use yasnippet. I have to mention that I was using Emacs a decade ago
when I switched to vim. For Org-mode I returned to GNU Emacs. Now I
have to re-orienteer again what is available and how it works :-)

yasnippet is superior to org-capture and it is not :-)

What is missing with yasnippet is a quick capture method which
includes the destination for the new item.

What is great with yasnippet is, that you can define certain input
values that have to be entered. Those values can be re-used. So when
you take a look at my snippet, you recognize $1, $2, ...

And you recognize that $1 and $2 a re-used. I do have to enter $1
only once and it is inserted multiple times - depending on the
occurrence of $1 in the rest.

For the :TYPE: property I let yasnippet ask me, whether it is a
person or a company I am entering now. Handy.

>> ,----
>> | # name : Org-contacts template for a person or a company
>> | # --
>> | ** $1 $2  :$1$2:
>> | :PROPERTIES:
>> | :TYPE: ${3:$$(yas/choose-value '("person" "company"))}
>> | :TITLE:
>> | :EMAIL: $4
>> | :URL:
>> | :MOBILE: 0043/
>> | :HOMEPHONE:
>> | :WORKPHONE:
>> | :PHONE:
>> | :COMPANY:
>> | :STREET:
>> | :POSTALCODE:
>> | :CITY:
>> | :COUNTRY: Österreich
>> | :END:
>> |
>> | first contact: $0
>> `----
>
> Thats what I thought, that a single :ADDRESS: property is not very
> usefull or practical. 

You loose meta information and semantic description.

With my method you can generate sparse trees with people living in a
certain street, town or such things.

> So there is not something like a nested property,
> each piece of information needs its own property. 

I do think so. But I am new to Org-mode and might still miss certain
details.

> I think your list is pretty complete and usable. 

Thanks.

>> My current tags for contacts are defined in the header of my
>> contacts.org:
>>
>> ,----[ first line of my contacts.org ]
>> | #+TAGS: job(j) friends(f) health(e) hotels(o) restaurants_bars(r) \
>> |         sport(s) students_TU(t) relatives(r) company(c)
>> `----
>
> [how do you produce this nice insertions in your email?]

As an Emacs user, you might not be delighted with my answer: I am
using mutt[3] as MUA[4]. There I can choose my editor freely. For
this purpose I am using vim 7.1. In vim I can use «boxes»[5]:

,----[ a line from my .vimrc ]
| map ;1 <ESC>:'<,'>!boxes -d boxquote
`----

But I also used to use Emacs for mail composing years ago and I do
recall a similar method there.

> Thanks for your help

You're welcome!

  3. http://www.mutt.org/
  4. http://en.wikipedia.org/wiki/Mail_user_agent
  5. http://boxes.thomasjensen.com/
  6. As I recognize here, the vim referencing mechanism seem to be
     confused by the $1 and $2 from above and started with [3]. Cute.
-- 
Karl Voit

Re: org-contacts: how to manage contacts; structure of meta data

[Thorsten]

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

> yasnippet is superior to org-capture and it is not :-)

So considering both and choosing the most suited for the problem might
be a good idea. 

> As an Emacs user, you might not be delighted with my answer: I am
> using mutt[3] as MUA[4]. There I can choose my editor freely. For
> this purpose I am using vim 7.1. 

Uhhh ... I shouldn't have asked ;)

> But I also used to use Emacs for mail composing years ago and I do
> recall a similar method there.

I'm sure there are, since I have seen some elaborated emails here, but
I'm not sure where to look for them. In the Gnus and the Message manuals
I did not find anything related so far. 

Cheers
-- 
Thorsten

Re: org-contacts: how to manage contacts; structure of meta data

[Wes Hardaker]

>>>>> On Thu, 6 Oct 2011 11:38:26 +0200, Karl Voit <devnull@Karl-Voit.at> said:

KV> yasnippet is superior to org-capture and it is not :-)

Well summarized :-).  The right tool for the right job definitely
matters here.

>> [how do you produce this nice insertions in your email?]

KV> As an Emacs user, you might not be delighted with my answer: I am
KV> using mutt[3] as MUA[4]. There I can choose my editor freely. For
KV> this purpose I am using vim 7.1. In vim I can use «boxes»[5]:

FYI, the emacs package you're looking for with similar functionality is
"rebox2" ( http://www.emacswiki.org/emacs/rebox2 )
-- 
Wes Hardaker                                     
My Pictures:  http://capturedonearth.com/
My Thoughts:  http://pontifications.hardakers.net/

Re: org-contacts: how to manage contacts; structure of meta data

[Thorsten]

Wes Hardaker <wjhns209@hardakers.net> writes:

> FYI, the emacs package you're looking for with similar functionality is
> "rebox2" ( http://www.emacswiki.org/emacs/rebox2 )

I knew - or at least expected - that there must be something like
that. Thanks for the link.
-- 
Thorsten