[BUG] htmlp and latexp

[BUG] htmlp and latexp

[Dan Davison]

Commit ed6d6760268 removed variables htmlp and latexp from
`org-export-preprocess-string'. Nothing wrong with that, but I think it
has broken export for those using org-special-blocks, which contains

#+begin_src emacs-lisp
(defvar htmlp)
(defvar latexp)
(defun org-special-blocks-make-special-cookies ()
  "Adds special cookies when #+begin_foo and #+end_foo tokens are
seen.  This is run after a few special cases are taken care of."
  (when (or htmlp latexp)
    (goto-char (point-min))
...
#+end_src

I think because those defvars lack a second argument, the variable is
not actually defined, i.e. they're not equivalent to (defvar htmlp nil).

IIuc htmlp and latexp occur occasionally in org code as somewhat
unofficial ways to test "am I in the middle of export?".

They also occur in org-exp-blocks, but in deprecated code, so not a
priority to fix.

Dan

Re: [BUG] htmlp and latexp

[Kieran Healy]

Yes, I ran into this earlier today as well ...

--
Kieran Healy :: http://www.kieranhealy.org

Re: [BUG] htmlp and latexp

[Christian Moe]

On 2/15/11 11:23 PM, Dan Davison wrote:
> Commit ed6d6760268 removed variables htmlp and latexp from
> `org-export-preprocess-string'. Nothing wrong with that, but I think it
> has broken export for those using org-special-blocks

Confirmed. Exporting the following example stops with a void-variable 
error in org-special-blocks-make-special-cookies.

--------------------

Some text.

#+begin_sidebar
   Some details left out of the main text.
#+end_sidebar

--------------------

(Incidentally -- org-special-blocks.el still says it's "not currently 
part of GNU Emacs", but as of Org-mode 7.4 it is, isn't it?)

CM

Re: [BUG] htmlp and latexp

[Bastien]

Hi Christian,

Christian Moe <mail@christianmoe.com> writes:

> (Incidentally -- org-special-blocks.el still says it's "not currently 
> part of GNU Emacs", but as of Org-mode 7.4 it is, isn't it?)

It's not yet part of GNU Emacs but it will be soon.

See commit 57144fea1ea6277b6ba39facae05715eee5751d6.

-- 
 Bastien

Re: [BUG] htmlp and latexp

[Bastien]

Hi Christian,

Christian Moe <mail@christianmoe.com> writes:

> (Incidentally -- org-special-blocks.el still says it's "not currently 
> part of GNU Emacs", but as of Org-mode 7.4 it is, isn't it?)

It's not yet part of GNU Emacs but it will be soon.

See commit 57144fea1ea6277b6ba39facae05715eee5751d6.

-- 
 Bastien

Re: [BUG] htmlp and latexp

[Bastien]

Hi Dan,

Dan Davison <dandavison7@gmail.com> writes:

> Commit ed6d6760268 removed variables htmlp and latexp from
> `org-export-preprocess-string'. Nothing wrong with that, but I think it
> has broken export for those using org-special-blocks, which contains
>
> #+begin_src emacs-lisp
> (defvar htmlp)
> (defvar latexp)
> (defun org-special-blocks-make-special-cookies ()
>   "Adds special cookies when #+begin_foo and #+end_foo tokens are
> seen.  This is run after a few special cases are taken care of."
>   (when (or htmlp latexp)
>     (goto-char (point-min))
> ...
> #+end_src

Fixed thanks.  

> IIuc htmlp and latexp occur occasionally in org code as somewhat
> unofficial ways to test "am I in the middle of export?".

Now there is a uniform and official (!) way of getting the backend the
user is currently exporting to: 

  (eq backend 'html)

...

> They also occur in org-exp-blocks, but in deprecated code, so not a
> priority to fix.

I fixed them there too.

Thanks for reporting this!

-- 
 Bastien

Re: [BUG] htmlp and latexp

[Dan Davison]

Christian Moe <mail@christianmoe.com> writes:

> On 2/15/11 11:23 PM, Dan Davison wrote:
>> Commit ed6d6760268 removed variables htmlp and latexp from
>> `org-export-preprocess-string'. Nothing wrong with that, but I think it
>> has broken export for those using org-special-blocks
>
> Confirmed. Exporting the following example stops with a void-variable 
> error in org-special-blocks-make-special-cookies.

I was about to make a temporary fix but I see Bastien's just fixed
it. Bastien -- should we give that `backend' variable a name within the
org-* namespace (and maybe defvar it in org.el?) so that it is a more
respectable way to test for "am I doing export now?"? (This issue just
came up in a separate babel thread).

Dan

>
> --------------------
>
> Some text.
>
> #+begin_sidebar
>    Some details left out of the main text.
> #+end_sidebar
>
> --------------------
>
> (Incidentally -- org-special-blocks.el still says it's "not currently 
> part of GNU Emacs", but as of Org-mode 7.4 it is, isn't it?)
>
> CM
>
>
> _______________________________________________
> Emacs-orgmode mailing list
> Please use `Reply All' to send replies to the list.
> Emacs-orgmode@gnu.org
> http://lists.gnu.org/mailman/listinfo/emacs-orgmode

Re: [BUG] htmlp and latexp

[Dan Davison]

Bastien <bastien.guerry@wikimedia.fr> writes:

> Hi Dan,
>
> Dan Davison <dandavison7@gmail.com> writes:
>
>> I was about to make a temporary fix but I see Bastien's just fixed
>> it. Bastien -- should we give that `backend' variable a name within the
>> org-* namespace (and maybe defvar it in org.el?) so that it is a more
>> respectable way to test for "am I doing export now?"? (This issue just
>> came up in a separate babel thread).
>
> backend is dynamically scoped, so I think there is no need to defvar it
> in org.el -- only in org-*-blocks.el, to avoid compilation warnings.
>
> Let me know if you think otherwise.

Hi Bastien,

I was thinking that if it were defvard in org.el, say as

(defvar org-export-current-backend nil)

then any code could use (null org-export-current-backend) to test
whether org is currently exporting or not. Otherwise, the code has to be
certain that it has been called during export if it is going to access
the variable.

For example, babel code can get called both during interactive
evaluation with C-c C-c, and also during export. It might be helpful for
that code to be able to inspect the value of org-export-current-backend
to see what the current context is.

Dan

>
> But it's a good idea to rename it to org-export-backend.  Can you do it?

Re: [BUG] htmlp and latexp

[Andreas Leha]

[-- Attachment #1.1.1: Type: text/plain, Size: 1450 bytes --]

Hi all,

to make this explicit:  I can not export to latex any more, but instead
I get
 Exporting to LaTeX...
 when: Symbol's value as variable is void: htmlp

Org HEAD
GNU emacs 23.2.1 (debian squeeze)

This is my failing org-file:
* Test
  test

- Andreas


Am 16.02.2011 11:03, schrieb Bastien:
> Hi Dan,
>
> Dan Davison <dandavison7@gmail.com> writes:
>
>   
>> Commit ed6d6760268 removed variables htmlp and latexp from
>> `org-export-preprocess-string'. Nothing wrong with that, but I think it
>> has broken export for those using org-special-blocks, which contains
>>
>> #+begin_src emacs-lisp
>> (defvar htmlp)
>> (defvar latexp)
>> (defun org-special-blocks-make-special-cookies ()
>>   "Adds special cookies when #+begin_foo and #+end_foo tokens are
>> seen.  This is run after a few special cases are taken care of."
>>   (when (or htmlp latexp)
>>     (goto-char (point-min))
>> ...
>> #+end_src
>>     
> Fixed thanks.  
>
>   
>> IIuc htmlp and latexp occur occasionally in org code as somewhat
>> unofficial ways to test "am I in the middle of export?".
>>     
> Now there is a uniform and official (!) way of getting the backend the
> user is currently exporting to: 
>
>   (eq backend 'html)
>
> ...
>
>   
>> They also occur in org-exp-blocks, but in deprecated code, so not a
>> priority to fix.
>>     
> I fixed them there too.
>
> Thanks for reporting this!
>
>   


[-- Attachment #1.1.2: andreas_leha.vcf --]
[-- Type: text/x-vcard, Size: 377 bytes --]

begin:vcard
fn:Andreas Leha
n:Leha;Andreas
org;quoted-printable:University Medicine G=C3=B6ttingen;Biostatistics/Bioinformatics (Dept. Medical Statistics)
adr;quoted-printable:;;Humboldallee 32;G=C3=B6ttingen;;37073;Germany
email;internet:andreas.leha@med.uni-goettingen.de
title:Dipl. Inf.
tel;work:+49 551 39-10710
tel;cell:+49 163 7662824
version:2.1
end:vcard


[-- Attachment #1.2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 6432 bytes --]

[-- Attachment #2: Type: text/plain, Size: 201 bytes --]

_______________________________________________
Emacs-orgmode mailing list
Please use `Reply All' to send replies to the list.
Emacs-orgmode@gnu.org
http://lists.gnu.org/mailman/listinfo/emacs-orgmode

Re: [BUG] htmlp and latexp

[Sébastien Vauban]

Hi Andreas,

Andreas Leha wrote:
> to make this explicit:  I can not export to latex any more, but instead
> I get
>  Exporting to LaTeX...
>  when: Symbol's value as variable is void: htmlp

I confirm you this has been fixed 1 or 2 days ago. Just git pull, and you're
gone.

Best regards,
  Seb

-- 
Sébastien Vauban


_______________________________________________
Emacs-orgmode mailing list
Please use `Reply All' to send replies to the list.
Emacs-orgmode-mXXj517/zsQ@public.gmane.org
http://lists.gnu.org/mailman/listinfo/emacs-orgmode

Re: [BUG] htmlp and latexp

[Jambunathan K]

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

> Hi all,
>
> to make this explicit:  I can not export to latex any more, but instead
> I get
>  Exporting to LaTeX...
>  when: Symbol's value as variable is void: htmlp
>
> Org HEAD
> GNU emacs 23.2.1 (debian squeeze)
>
> This is my failing org-file:
> * Test
>   test
>

Have you tried recent head?  
Enable stacktrace: M-x toggle-debug-on-error
Also include M-x org-version? 

Jambunathan K.


> - Andreas
>
>
> Am 16.02.2011 11:03, schrieb Bastien:
>> Hi Dan,
>>
>> Dan Davison <dandavison7@gmail.com> writes:
>>
>>   
>>> Commit ed6d6760268 removed variables htmlp and latexp from
>>> `org-export-preprocess-string'. Nothing wrong with that, but I think it
>>> has broken export for those using org-special-blocks, which contains
>>>
>>> #+begin_src emacs-lisp
>>> (defvar htmlp)
>>> (defvar latexp)
>>> (defun org-special-blocks-make-special-cookies ()
>>>   "Adds special cookies when #+begin_foo and #+end_foo tokens are
>>> seen.  This is run after a few special cases are taken care of."
>>>   (when (or htmlp latexp)
>>>     (goto-char (point-min))
>>> ...
>>> #+end_src
>>>     
>> Fixed thanks.  
>>
>>   
>>> IIuc htmlp and latexp occur occasionally in org code as somewhat
>>> unofficial ways to test "am I in the middle of export?".
>>>     
>> Now there is a uniform and official (!) way of getting the backend the
>> user is currently exporting to: 
>>
>>   (eq backend 'html)
>>
>> ...
>>
>>   
>>> They also occur in org-exp-blocks, but in deprecated code, so not a
>>> priority to fix.
>>>     
>> I fixed them there too.
>>
>> Thanks for reporting this!
>>
>>   
>
>
> _______________________________________________
> Emacs-orgmode mailing list
> Please use `Reply All' to send replies to the list.
> Emacs-orgmode@gnu.org
> http://lists.gnu.org/mailman/listinfo/emacs-orgmode

Re: [BUG] htmlp and latexp

[Bastien]

Hi Dan,

Dan Davison <dandavison7@gmail.com> writes:

> I was thinking that if it were defvard in org.el, say as
>
> (defvar org-export-current-backend nil)
>
> then any code could use (null org-export-current-backend) to test
> whether org is currently exporting or not. Otherwise, the code has to be
> certain that it has been called during export if it is going to access
> the variable.

I agree this would be a good idea.  If you have time, please send a
patch.  Otherwise, I just put it [somewhere] in my own todo list...

Thanks!

-- 
 Bastien