Re: Re: [BUG] nested blocks aren't protected on export

[Carsten Dominik <carsten.dominik@gmail.com>]

Hi Eric,

On Jun 28, 2010, at 9:59 PM, Eric Schulte wrote:

> Carsten Dominik <carsten.dominik@gmail.com> writes:
>
> [...]
>>
>> For now I don't see anything wrong with this.  The fastest
>> way to find out is to put it into master and watch the list :-)
>> Please go ahead and do that.
>>
>
> I think this patch may have uncovered a bug in the export  
> preprocessing
> which was previously being covered up by the over-eager comment  
> pruning.
>
> I've now got another patch [1] which fixes the bug in
> `org-export-mark-blockquote-verse-center' which was leaving # 
> +end_quote
> lines littered in the pre-processed files.
>
> If this looks good to you please do apply it, I'll run it locally in  
> the
> meantime to see if any unexpected behavior develops.
>
> Thanks -- Eric
>
>>
>> Thanks.
>>
>> - Carsten
>
> Footnotes:
> [1]  From 251647d34f599dd025bb08376c0a3eabac604c9d Mon Sep 17  
> 00:00:00 2001
> From: Eric Schulte <schulte.eric@gmail.com>
> Date: Mon, 28 Jun 2010 12:52:44 -0700
> Subject: [PATCH] org-exp: ensure that #+end_quote lines are removed
>
> * lisp/org-exp.el (org-export-mark-blockquote-verse-center): ensure
>  that the end_quote line is removed during buffer preprocessing
> ---
> lisp/org-exp.el |    3 ++-
> 1 files changed, 2 insertions(+), 1 deletions(-)
>
> diff --git a/lisp/org-exp.el b/lisp/org-exp.el
> index f0c11d6..1cd8fc2 100644
> --- a/lisp/org-exp.el
> +++ b/lisp/org-exp.el
> @@ -1569,7 +1569,8 @@ These special cookies will later be  
> interpreted by the backend."
> 	    beg1 (1+ (match-end 0)))
>       (when (re-search-forward (concat "^[ \t]*#\\+end_" type "\ 
> \>.*") nil t)
> 	(setq end (+ (point-at-eol) (if (looking-at "\n$") 1 0))
> -	      end1 (1- (match-beginning 0)))
> +	      end1 (save-excursion (goto-char (1- (match-beginning 0)))
> +				   (point-at-bol)))
> 	(setq content (org-remove-indentation (buffer-substring beg1 end1)))
> 	(setq content (concat "ORG-" (upcase t1) "-START\n"
> 			      content "\n"
> -- 
> 1.7.0.4

I think the better fix woud be

diff --git a/lisp/org-exp.el b/lisp/org-exp.el
index f0c11d6..23fb07b 100644
--- a/lisp/org-exp.el
+++ b/lisp/org-exp.el
@@ -1568,8 +1568,8 @@ These special cookies will later be interpreted  
by the bac
        (setq beg (match-beginning 0)
             beg1 (1+ (match-end 0)))
        (when (re-search-forward (concat "^[ \t]*#\\+end_" type "\ 
\>.*") nil t)
-       (setq end (+ (point-at-eol) (if (looking-at "\n$") 1 0))
-             end1 (1- (match-beginning 0)))
+       (setq end1 (1- (match-beginning 0))
+             end (+ (point-at-eol) (if (looking-at "\n$") 1 0)))
         (setq content (org-remove-indentation (buffer-substring beg1  
end1)))
         (setq content (concat "ORG-" (upcase t1) "-START\n"
                               content "\n"


The reason why this fails seems to be because of the looking-at.
When this code was first written, (match-beginning 0) referred to
the earlier regexp match, but the looking-at, if it matches,
changes the meaning of that form.

Could you please try if my patch works as well, and if yes apply it?

Thanks.

- Carsten



>

- Carsten