Regression in current trunk?

Regression in current trunk?

[Angelo Graziosi]

The following happens on Cygwin, but perhaps it is general.

After that Eli fixed the status line [1], I did a new build and, really,
now the status line is OK.

But... I observe that Emacs is very slowly when one visit a big text
file (about 12M, the log of Cernlib build): it seems to hang when one
tries to move around up, down etc.. : for example, typing the down key,
it take about 1 second to move the cursor on the next line. It is almost
impossible to search (C-s) words.

If in addition I start also another Emacs, built a few days ago
(2008.08.22), and visit the same big file, it is very fast in moving
around: no delay in moving the cursor etc. To search words with C-s is
immediate.

Does it look as a regression?


Cheers,
    Angelo.

---
[1] http://lists.gnu.org/archive/html/emacs-devel/2008-08/msg01305.html

Re: Regression in current trunk?

[Angelo Graziosi]

Angelo Graziosi ha scritto:
> The following happens on Cygwin, but perhaps it is general.
> 

It happens also on GNU/Linux Kubuntu 8.04.1, and also with the trunk of
few minutes ago...

    Angelo


> After that Eli fixed the status line [1], I did a new build and, really,
> now the status line is OK.
> 
> But... I observe that Emacs is very slowly when one visit a big text
> file (about 12M, the log of Cernlib build): it seems to hang when one
> tries to move around up, down etc.. : for example, typing the down key,
> it take about 1 second to move the cursor on the next line. It is almost
> impossible to search (C-s) words.
> 
> If in addition I start also another Emacs, built a few days ago
> (2008.08.22), and visit the same big file, it is very fast in moving
> around: no delay in moving the cursor etc. To search words with C-s is
> immediate.
> 
> Does it look as a regression?
> 
> 
> Cheers,
>    Angelo.
> 
> ---
> [1] http://lists.gnu.org/archive/html/emacs-devel/2008-08/msg01305.html
> 

Re: Regression in current trunk?

[Angelo Graziosi]

[-- Attachment #1: Type: text/plain, Size: 1803 bytes --]

Well, with a binary search, here there are a few results:


cvs up -dP -D "20080829 09:50" builds and works fine with big file

cvs up -dP -D "20080829 09:55" fails to build

cvs up -dP -D "20080829 10:00" builds and shows the regression discussed
                                here


These changes seem the cause:

2008-08-29 Kenichi Handa <xxx@xxx>

        These changes are to re-implement the automatic composition so
        that it doesn't use text properties.

        * Makefile.in (ftfont.o): Depend on composite.h.
        (composite.o): Depend dispextern.h, font.h, frame, and window.h.

        * character.h (Vunicode_category_table): Extern it.
[...]


Attached the source differences.


Cheers,
    Angelo.

Angelo Graziosi ha scritto:
> Angelo Graziosi ha scritto:
>> The following happens on Cygwin, but perhaps it is general.
>>
> 
> It happens also on GNU/Linux Kubuntu 8.04.1, and also with the trunk of
> few minutes ago...
> 
>    Angelo
> 
> 
>> After that Eli fixed the status line [1], I did a new build and, really,
>> now the status line is OK.
>>
>> But... I observe that Emacs is very slowly when one visit a big text
>> file (about 12M, the log of Cernlib build): it seems to hang when one
>> tries to move around up, down etc.. : for example, typing the down key,
>> it take about 1 second to move the cursor on the next line. It is almost
>> impossible to search (C-s) words.
>>
>> If in addition I start also another Emacs, built a few days ago
>> (2008.08.22), and visit the same big file, it is very fast in moving
>> around: no delay in moving the cursor etc. To search words with C-s is
>> immediate.
>>
>> Does it look as a regression?
>>
>>
>> Cheers,
>>    Angelo.
>>
>> ---
>> [1] http://lists.gnu.org/archive/html/emacs-devel/2008-08/msg01305.html
>>
> 
> 


[-- Attachment #2: emacs-src.diff.bz2 --]
[-- Type: application/octet-stream, Size: 30941 bytes --]