bug#1952: 23.0.60[W32-Mingw]; opening etc/HELLO is slower than before

bug#1952: 23.0.60[W32-Mingw]; opening etc/HELLO is slower than before

[Kevin Yu]

Hi,

    C-h h. It takes about 3~4 seconds to render the buffer. And if you
try beginning-of-buffer/end-of-buffer,
you will notice that it's a little lag to refresh.
    Here's hardware configuration of my PC:
          2.6GHz CPU
          512M MEM

Best wishes


               Kevin

bug#1952: 23.0.60[W32-Mingw]; opening etc/HELLO is slower than before

[Jason Rumney]

"slower than before" means what exactly? Can you pinpoint the change 
that caused the slowdown? Or are you comparing with Emacs 22.3?

If the latter, then at least part of the slowdown can be attributed to 
the fact that Emacs 23 is finding all the fonts it can to display the 
characters in etc/HELLO correctly, while Emacs 22.3 displays about half 
the languages as empty boxes, and another couple of languages as the 
font's "undefined" glyph.

bug#1952: 23.0.60[W32-Mingw]; opening etc/HELLO is slower than before

[Juanma Barranquero]

On Mon, Jan 19, 2009 at 15:14, Jason Rumney <jasonr@gnu.org> wrote:

> "slower than before" means what exactly? Can you pinpoint the change that
> caused the slowdown? Or are you comparing with Emacs 22.3?
>
> If the latter, then at least part of the slowdown can be attributed to the
> fact that Emacs 23 is finding all the fonts it can to display the characters
> in etc/HELLO correctly, while Emacs 22.3 displays about half the languages
> as empty boxes, and another couple of languages as the font's "undefined"
> glyph.

I think Kevin Yu is right. Moving around in etc/HELLO, even with a lot
of installed fonts, was quite fast until this change:

2009-01-09  Jason Rumney  <jasonr@gnu.org>

        * w32font.c (add_font_entity_to_list): Don't report unknown
        Windows charset as any unrecognized registry.  (Bug#1548)
        Only report Unicode Plane 2 fonts as unicode-sip.

Now, moving in etc/HELLO is unbearably slow. I'm talking of typing
<down> and waiting about two seconds for the cursor to move.

    Juanma

bug#1952: 23.0.60[W32-Mingw]; opening etc/HELLO is slower than before

[Jason Rumney]

Juanma Barranquero wrote:
> I think Kevin Yu is right. Moving around in etc/HELLO, even with a lot
> of installed fonts, was quite fast until this change:
>
> 2009-01-09  Jason Rumney  <jasonr@gnu.org>
>
>         * w32font.c (add_font_entity_to_list): Don't report unknown
>         Windows charset as any unrecognized registry.  (Bug#1548)
>         Only report Unicode Plane 2 fonts as unicode-sip.
>
> Now, moving in etc/HELLO is unbearably slow. I'm talking of typing
> <down> and waiting about two seconds for the cursor to move.
>   

I don't see any slowness when scrolling, but it may be because I do not 
have a lot of fonts installed compared with some people, or it may be 
because I have fonts installed for all the languages in HELLO, so there 
are no failures (which I suspect may not get cached).

One of the changes made at the same time as the above (which I 
apparently forgot to add a ChangeLog entry for) was to fix the return 
value of add_one_font_entity_to_list. This function was supposed to 
abort the font enumeration as soon as a match was found. But previously 
the logic was wrong so it was aborting if the first font did not match 
the given font-spec. So in the majority of cases, it would only 
enumerate one font and return nil, only if the first font matched would 
it go through every font in the system and return all the fonts that 
matched. After the fix, it will go through all the fonts for characters 
that do not have a matching font installed, and will go through enough 
fonts to find a match for others.

Since there was no visible bug previously, I wonder what use this 
function is, and whether it is just wasting time looking for a match 
that we already have.

bug#1952: 23.0.60[W32-Mingw]; opening etc/HELLO is slower than before

[Juanma Barranquero]

On Tue, Jan 20, 2009 at 13:52, Jason Rumney <jasonr@gnu.org> wrote:

> This function was supposed to abort the font
> enumeration as soon as a match was found. But previously the logic was wrong
> so it was aborting if the first font did not match the given font-spec. So
> in the majority of cases, it would only enumerate one font and return nil,
> only if the first font matched would it go through every font in the system
> and return all the fonts that matched. After the fix, it will go through all
> the fonts for characters that do not have a matching font installed, and
> will go through enough fonts to find a match for others.

The change that causes the slowdown in my system is not the one to
add_one_font_entity_to_list, but this one to add_font_entity_to_list:

@@ -1449,6 +1449,5 @@
                    && !EQ (spec_charset, Qunicode_bmp)
                    && !EQ (spec_charset, Qunicode_sip)
-                   && match_data->pattern.lfCharSet == DEFAULT_CHARSET
-                   && logical_font->elfLogFont.lfCharSet != DEFAULT_CHARSET)
+                   && match_data->pattern.lfCharSet == DEFAULT_CHARSET)
             return 0;

Reverting this bit eliminates the problem.

    Juanma

bug#1952: 23.0.60[W32-Mingw]; opening etc/HELLO is slower than before

[Jason Rumney]

Juanma Barranquero wrote:
> The change that causes the slowdown in my system is not the one to
> add_one_font_entity_to_list, but this one to add_font_entity_to_list:
>
> @@ -1449,6 +1449,5 @@
>                     && !EQ (spec_charset, Qunicode_bmp)
>                     && !EQ (spec_charset, Qunicode_sip)
> -                   && match_data->pattern.lfCharSet == DEFAULT_CHARSET
> -                   && logical_font->elfLogFont.lfCharSet != DEFAULT_CHARSET)
> +                   && match_data->pattern.lfCharSet == DEFAULT_CHARSET)
>              return 0;
>   
The earlier code never returned 0, always 1, so I don't think that is 
the actual change.

> Reverting this bit eliminates the problem.
>   

Strange, I'd expect that to speed things up, since the new code aborts a 
search that will never find a match where the old code would go through 
all the fonts, returning false matches that would fail later tests.


Can you debug this to see why the earlier version is faster?

bug#1952: 23.0.60[W32-Mingw]; opening etc/HELLO is slower than before

[Juanma Barranquero]

On Tue, Jan 20, 2009 at 17:05, Jason Rumney <jasonr@gnu.org> wrote:

> The earlier code never returned 0, always 1, so I don't think that is the
> actual change.

This is not the *actual* change, it's part of the change you
committed. But reverting just that bit is enough for Emacs to behave
as it did before w32font.c rev 1.67.

> Can you debug this to see why the earlier version is faster?

Sure, I'll try. Any pointer about what/where to debug is very welcome.

    Juanma

bug#1952: 23.0.60[W32-Mingw]; opening etc/HELLO is slower than before

[Kevin Yu]

On Mon, Jan 19, 2009 at 10:14 PM, Jason Rumney <jasonr@gnu.org> wrote:
> "slower than before" means what exactly? Can you pinpoint the change that
> caused the slowdown? Or are you comparing with Emacs 22.3?

comparing with emacs 23.


>
> If the latter, then at least part of the slowdown can be attributed to the
> fact that Emacs 23 is finding all the fonts it can to display the characters
> in etc/HELLO correctly, while Emacs 22.3 displays about half the languages
> as empty boxes, and another couple of languages as the font's "undefined"
> glyph.

But when I'm using beginning-of-buffer/end-of-buffer repeatedly, It
still lags. Is emacs
finding fonts at this situation?

>
>
>

bug#1952: 23.0.60[W32-Mingw]; opening etc/HELLO is slower than before

[Jason Rumney]

Kevin Yu wrote:
> But when I'm using beginning-of-buffer/end-of-buffer repeatedly, It
> still lags. Is emacs
> finding fonts at this situation?
>   

It shouldn't be, but perhaps we are not caching failures.

Can you please try the following:

M-: (setq font-log nil)
C-h h
C-End
C-Home
M-x font-show-log

and paste the result in a reply to this.

bug#1952: marked as done (23.0.60[W32-Mingw];opening etc/HELLO is slower than before)

[Emacs bug Tracking System]

[-- Attachment #1: Type: text/plain, Size: 916 bytes --]


Your message dated Sat, 24 Jan 2009 17:44:13 -0500
with message-id <874ozowchu.fsf@cyd.mit.edu>
and subject line Re: bug#1990: 23.0.60; Emacs trying too hard to find fonts that don't  exist
has caused the Emacs bug report #1952,
regarding 23.0.60[W32-Mingw];opening etc/HELLO is slower than before
to be marked as done.

This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
bug report if necessary, and/or fix the problem forthwith.

(NB: If you are a system administrator and have no idea what this
message is talking about, this may indicate a serious mail system
misconfiguration somewhere. Please contact owner@emacsbugs.donarmstrong.com
immediately.)


-- 
1952: http://emacsbugs.donarmstrong.com/cgi-bin/bugreport.cgi?bug=1952
Emacs Bug Tracking System
Contact owner@emacsbugs.donarmstrong.com with problems

[-- Attachment #2: Type: message/rfc822, Size: 3153 bytes --]

From: Kevin Yu <yujie052@gmail.com>
To: emacs-pretest-bug <emacs-pretest-bug@gnu.org>
Subject: 23.0.60[W32-Mingw];opening etc/HELLO is slower than before
Date: Mon, 19 Jan 2009 10:40:09 +0800
Message-ID: <42b562540901181840v1cfc73bbxec292b40aa4a60aa@mail.gmail.com>

Hi,

    C-h h. It takes about 3~4 seconds to render the buffer. And if you
try beginning-of-buffer/end-of-buffer,
you will notice that it's a little lag to refresh.
    Here's hardware configuration of my PC:
          2.6GHz CPU
          512M MEM

Best wishes


               Kevin



[-- Attachment #3: Type: message/rfc822, Size: 1948 bytes --]

From: Chong Yidong <cyd@stupidchicken.com>
To: Juanma Barranquero <lekktu@gmail.com>
Cc: Jason Rumney <jasonr@gnu.org>, Kenichi Handa <handa@m17n.org>, 1990-done@emacsbugs.donarmstrong.com, 1952-done@emacsbugs.donarmstrong.com
Subject: Re: bug#1990: 23.0.60; Emacs trying too hard to find fonts that don't  exist
Date: Sat, 24 Jan 2009 17:44:13 -0500
Message-ID: <874ozowchu.fsf@cyd.mit.edu>

Juanma Barranquero <lekktu@gmail.com> writes:

>> That log message is emitted just before calling fontset_font (which is
>> where we would check whether fonts are known not to exist).
>
> OK.
>
> Then, please install your change. It's a big improvement.

I have installed a similar patch in CVS.

bug#1990: 23.0.60; Emacs trying too hard to find fonts that don't exist

[Jason Rumney]

Please write in English if possible, because the Emacs maintainers
usually do not have translators to read other languages for them.

Your bug report will be posted to the emacs-pretest-bug@gnu.org mailing 
list.

Please describe exactly what actions triggered the bug
and the precise symptoms of the bug:

If fonts are not installed for some characters, then Emacs keeps trying 
to find fonts for them every time those characters are redisplayed. The 
following excerpt from font-log is with the xft font backend on 
GNU/Linux where the search is quick, thus not noticeable, but on Windows 
the search for fonts is expensive, so redisplay takes a couple of 
seconds. I have cut the end off, as the full font-log is long, but you 
can see from the first two font-for entries that Emacs is trying twice 
for the same character (and it tries again many times further down the log).

font for: (4121 . mule-unicode-0100-24ff)
list: -mutt-clearlyu-*-iso10646-1
list: -gnu-unifont-*-iso10646-1
list: -bitstream-Bitstream Vera Sans Mono-*-gb18030*-*
list: -*-Bitstream Vera Sans Mono-*-gb18030*-*
list: -bitstream-*-gb18030*-*
list: -*-gb18030*-*
list: -bitstream-Bitstream Vera Sans Mono-*-ksc5601.1987*-*
list: -*-Bitstream Vera Sans Mono-*-ksc5601.1987*-*
list: -bitstream-*-ksc5601.1987*-*
list: -*-ksc5601.1987*-*
list: -bitstream-Bitstream Vera Sans Mono-*-cns11643.1992-1
list: -*-Bitstream Vera Sans Mono-*-cns11643.1992-1
list: -bitstream-*-cns11643.1992-1
list: -*-cns11643.1992-1
list: -bitstream-Bitstream Vera Sans Mono-*-cns11643.1992-4
list: -*-Bitstream Vera Sans Mono-*-cns11643.1992-4
list: -bitstream-*-cns11643.1992-4
list: -*-cns11643.1992-4
list: -bitstream-Bitstream Vera Sans Mono-*-cns11643.1992-5
list: -*-Bitstream Vera Sans Mono-*-cns11643.1992-5
list: -bitstream-*-cns11643.1992-5
list: -*-cns11643.1992-5
list: -bitstream-Bitstream Vera Sans Mono-*-cns11643.1992-6
list: -*-Bitstream Vera Sans Mono-*-cns11643.1992-6
list: -bitstream-*-cns11643.1992-6
list: -*-cns11643.1992-6
list: -bitstream-Bitstream Vera Sans Mono-*-cns11643.1992-7
list: -*-Bitstream Vera Sans Mono-*-cns11643.1992-7
list: -bitstream-*-cns11643.1992-7
list: -*-cns11643.1992-7
font for: (4121 . mule-unicode-0100-24ff)
list: -mutt-clearlyu-*-iso10646-1
list: -gnu-unifont-*-iso10646-1
list: -bitstream-Bitstream Vera Sans Mono-*-gb18030*-*
list: -*-Bitstream Vera Sans Mono-*-gb18030*-*
list: -bitstream-*-gb18030*-*
list: -*-gb18030*-*
list: -bitstream-Bitstream Vera Sans Mono-*-ksc5601.1987*-*
list: -*-Bitstream Vera Sans Mono-*-ksc5601.1987*-*
list: -bitstream-*-ksc5601.1987*-*
list: -*-ksc5601.1987*-*
list: -bitstream-Bitstream Vera Sans Mono-*-cns11643.1992-1
list: -*-Bitstream Vera Sans Mono-*-cns11643.1992-1
list: -bitstream-*-cns11643.1992-1
list: -*-cns11643.1992-1
list: -bitstream-Bitstream Vera Sans Mono-*-cns11643.1992-4
list: -*-Bitstream Vera Sans Mono-*-cns11643.1992-4
list: -bitstream-*-cns11643.1992-4
list: -*-cns11643.1992-4
list: -bitstream-Bitstream Vera Sans Mono-*-cns11643.1992-5
list: -*-Bitstream Vera Sans Mono-*-cns11643.1992-5
list: -bitstream-*-cns11643.1992-5
list: -*-cns11643.1992-5
list: -bitstream-Bitstream Vera Sans Mono-*-cns11643.1992-6
list: -*-Bitstream Vera Sans Mono-*-cns11643.1992-6
list: -bitstream-*-cns11643.1992-6
list: -*-cns11643.1992-6
list: -bitstream-Bitstream Vera Sans Mono-*-cns11643.1992-7
list: -*-Bitstream Vera Sans Mono-*-cns11643.1992-7
list: -bitstream-*-cns11643.1992-7
list: -*-cns11643.1992-7


To reproduce: C-h h
On my system, there are boxes displayed for Sinhala, Tibetan, Khmer and 
Burmese.
Scroll down so that the first set of boxes disappears just off the top 
of the window.
M-: (setq font-log nil)
Scroll back up
M-x font-show-log

If Emacs crashed, and you have the Emacs process in the gdb debugger,
please include the output from the following gdb commands:
    `bt full' and `xbacktrace'.
If you would like to further debug the crash, please read the file
/home/jasonr/emacs/etc/DEBUG for instructions.


In GNU Emacs 23.0.60.1 (x86_64-unknown-linux-gnu, GTK+ Version 2.12.11)
 of 2008-12-28 on wanchan.jasonrumney.net
Windowing system distributor `Colin Harrison', version 11.0.60900031
Important settings:
  value of $LC_ALL: nil
  value of $LC_COLLATE: nil
  value of $LC_CTYPE: nil
  value of $LC_MESSAGES: nil
  value of $LC_MONETARY: nil
  value of $LC_NUMERIC: nil
  value of $LC_TIME: nil
  value of $LANG: en_GB.UTF-8
  value of $XMODIFIERS: nil
  locale-coding-system: utf-8-unix
  default-enable-multibyte-characters: t

Major mode: Help

Minor modes in effect:
  tooltip-mode: t
  tool-bar-mode: t
  mouse-wheel-mode: t
  menu-bar-mode: t
  file-name-shadow-mode: t
  global-font-lock-mode: t
  font-lock-mode: t
  blink-cursor-mode: t
  global-auto-composition-mode: t
  auto-composition-mode: t
  auto-encryption-mode: t
  auto-compression-mode: t
  line-number-mode: t
  transient-mark-mode: t
  view-mode: t

Recent input:
<help-echo> <help-echo> <help-echo> <help-echo> <escape>
x C-g C-h h <help-echo> <next> <prior> <down> <down>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <escape> : ( s e t q
SPC f o n t - l o g SPC n i l ) <return> <prior> <escape>
x f o n t - s h <tab> <return> <down-mouse-1> <mouse-movement>
<mouse-movement> <mouse-movement> <mouse-movement>
<mouse-movement> <drag-mouse-1> <down-mouse-1> <mouse-1>
<prior> <prior> <up> <up> <up> <up> <up> <up> <up>
<up> <up> <up> <up> <up> <up> <up> <up> <up> <up> C-@
<down> <down> <down> <down> <down> <down> <down> <down>
<down> <down> <down> <down> <down> <down> <down> <down>
<down> <down> <down> <down> <down> <down> <down> <down>
<down> <down> <down> <down> <down> <down> <down> <down>
<down> <down> <down> <down> <down> <down> <down> <down>
<down> <down> <down> <down> <down> <down> <down> <down>
<down> <down> <down> <down> <down> <down> <down> <down>
<down> <down> <down> <down> <down> <down> <escape>
w <help-echo> <help-echo> <help-echo> <help-echo> <escape>
x r e p o r t - e m a <tab> <return>

Recent messages:
For information about GNU Emacs and the GNU system, type C-h C-a.
Quit
Loading vc-cvs...done
View mode: type C-h for help, h for commands, q to quit.
nil
Mark activated
Saved text from "font for: (4121 . mule-unicode-0100-24ff"

bug#1990: 23.0.60; Emacs trying too hard to find fonts that don't exist

[Juanma Barranquero]

On Thu, Jan 22, 2009 at 14:22, Jason Rumney <jasonr@gnu.org> wrote:

> If fonts are not installed for some characters, then Emacs keeps trying to
> find fonts for them every time those characters are redisplayed.

Is this bug related to #1952, "23.0.60[W32-Mingw];opening etc/HELLO is
slower than before"?

    Juanma

bug#1990: 23.0.60; Emacs trying too hard to find fonts that don't exist

[Jason Rumney]

Juanma Barranquero wrote:
> On Thu, Jan 22, 2009 at 14:22, Jason Rumney <jasonr@gnu.org> wrote:
>   
>> If fonts are not installed for some characters, then Emacs keeps trying to
>> find fonts for them every time those characters are redisplayed.
>>     
>
> Is this bug related to #1952, "23.0.60[W32-Mingw];opening etc/HELLO is
> slower than before"?
>   

Probably.

bug#1990: marked as done (23.0.60; Emacs trying too hard to find fonts that don't exist)

[Emacs bug Tracking System]

[-- Attachment #1: Type: text/plain, Size: 919 bytes --]


Your message dated Sat, 24 Jan 2009 17:44:13 -0500
with message-id <874ozowchu.fsf@cyd.mit.edu>
and subject line Re: bug#1990: 23.0.60; Emacs trying too hard to find fonts that don't  exist
has caused the Emacs bug report #1990,
regarding 23.0.60; Emacs trying too hard to find fonts that don't exist
to be marked as done.

This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
bug report if necessary, and/or fix the problem forthwith.

(NB: If you are a system administrator and have no idea what this
message is talking about, this may indicate a serious mail system
misconfiguration somewhere. Please contact owner@emacsbugs.donarmstrong.com
immediately.)


-- 
1990: http://emacsbugs.donarmstrong.com/cgi-bin/bugreport.cgi?bug=1990
Emacs Bug Tracking System
Contact owner@emacsbugs.donarmstrong.com with problems

[-- Attachment #2: Type: message/rfc822, Size: 10429 bytes --]

From: Jason Rumney <jasonr@gnu.org>
To: emacs-pretest-bug <emacs-pretest-bug@gnu.org>
Subject: 23.0.60; Emacs trying too hard to find fonts that don't exist
Date: Thu, 22 Jan 2009 21:22:32 +0800
Message-ID: <49787318.20708@gnu.org>

Please write in English if possible, because the Emacs maintainers
usually do not have translators to read other languages for them.

Your bug report will be posted to the emacs-pretest-bug@gnu.org mailing 
list.

Please describe exactly what actions triggered the bug
and the precise symptoms of the bug:

If fonts are not installed for some characters, then Emacs keeps trying 
to find fonts for them every time those characters are redisplayed. The 
following excerpt from font-log is with the xft font backend on 
GNU/Linux where the search is quick, thus not noticeable, but on Windows 
the search for fonts is expensive, so redisplay takes a couple of 
seconds. I have cut the end off, as the full font-log is long, but you 
can see from the first two font-for entries that Emacs is trying twice 
for the same character (and it tries again many times further down the log).

font for: (4121 . mule-unicode-0100-24ff)
list: -mutt-clearlyu-*-iso10646-1
list: -gnu-unifont-*-iso10646-1
list: -bitstream-Bitstream Vera Sans Mono-*-gb18030*-*
list: -*-Bitstream Vera Sans Mono-*-gb18030*-*
list: -bitstream-*-gb18030*-*
list: -*-gb18030*-*
list: -bitstream-Bitstream Vera Sans Mono-*-ksc5601.1987*-*
list: -*-Bitstream Vera Sans Mono-*-ksc5601.1987*-*
list: -bitstream-*-ksc5601.1987*-*
list: -*-ksc5601.1987*-*
list: -bitstream-Bitstream Vera Sans Mono-*-cns11643.1992-1
list: -*-Bitstream Vera Sans Mono-*-cns11643.1992-1
list: -bitstream-*-cns11643.1992-1
list: -*-cns11643.1992-1
list: -bitstream-Bitstream Vera Sans Mono-*-cns11643.1992-4
list: -*-Bitstream Vera Sans Mono-*-cns11643.1992-4
list: -bitstream-*-cns11643.1992-4
list: -*-cns11643.1992-4
list: -bitstream-Bitstream Vera Sans Mono-*-cns11643.1992-5
list: -*-Bitstream Vera Sans Mono-*-cns11643.1992-5
list: -bitstream-*-cns11643.1992-5
list: -*-cns11643.1992-5
list: -bitstream-Bitstream Vera Sans Mono-*-cns11643.1992-6
list: -*-Bitstream Vera Sans Mono-*-cns11643.1992-6
list: -bitstream-*-cns11643.1992-6
list: -*-cns11643.1992-6
list: -bitstream-Bitstream Vera Sans Mono-*-cns11643.1992-7
list: -*-Bitstream Vera Sans Mono-*-cns11643.1992-7
list: -bitstream-*-cns11643.1992-7
list: -*-cns11643.1992-7
font for: (4121 . mule-unicode-0100-24ff)
list: -mutt-clearlyu-*-iso10646-1
list: -gnu-unifont-*-iso10646-1
list: -bitstream-Bitstream Vera Sans Mono-*-gb18030*-*
list: -*-Bitstream Vera Sans Mono-*-gb18030*-*
list: -bitstream-*-gb18030*-*
list: -*-gb18030*-*
list: -bitstream-Bitstream Vera Sans Mono-*-ksc5601.1987*-*
list: -*-Bitstream Vera Sans Mono-*-ksc5601.1987*-*
list: -bitstream-*-ksc5601.1987*-*
list: -*-ksc5601.1987*-*
list: -bitstream-Bitstream Vera Sans Mono-*-cns11643.1992-1
list: -*-Bitstream Vera Sans Mono-*-cns11643.1992-1
list: -bitstream-*-cns11643.1992-1
list: -*-cns11643.1992-1
list: -bitstream-Bitstream Vera Sans Mono-*-cns11643.1992-4
list: -*-Bitstream Vera Sans Mono-*-cns11643.1992-4
list: -bitstream-*-cns11643.1992-4
list: -*-cns11643.1992-4
list: -bitstream-Bitstream Vera Sans Mono-*-cns11643.1992-5
list: -*-Bitstream Vera Sans Mono-*-cns11643.1992-5
list: -bitstream-*-cns11643.1992-5
list: -*-cns11643.1992-5
list: -bitstream-Bitstream Vera Sans Mono-*-cns11643.1992-6
list: -*-Bitstream Vera Sans Mono-*-cns11643.1992-6
list: -bitstream-*-cns11643.1992-6
list: -*-cns11643.1992-6
list: -bitstream-Bitstream Vera Sans Mono-*-cns11643.1992-7
list: -*-Bitstream Vera Sans Mono-*-cns11643.1992-7
list: -bitstream-*-cns11643.1992-7
list: -*-cns11643.1992-7


To reproduce: C-h h
On my system, there are boxes displayed for Sinhala, Tibetan, Khmer and 
Burmese.
Scroll down so that the first set of boxes disappears just off the top 
of the window.
M-: (setq font-log nil)
Scroll back up
M-x font-show-log

If Emacs crashed, and you have the Emacs process in the gdb debugger,
please include the output from the following gdb commands:
    `bt full' and `xbacktrace'.
If you would like to further debug the crash, please read the file
/home/jasonr/emacs/etc/DEBUG for instructions.


In GNU Emacs 23.0.60.1 (x86_64-unknown-linux-gnu, GTK+ Version 2.12.11)
 of 2008-12-28 on wanchan.jasonrumney.net
Windowing system distributor `Colin Harrison', version 11.0.60900031
Important settings:
  value of $LC_ALL: nil
  value of $LC_COLLATE: nil
  value of $LC_CTYPE: nil
  value of $LC_MESSAGES: nil
  value of $LC_MONETARY: nil
  value of $LC_NUMERIC: nil
  value of $LC_TIME: nil
  value of $LANG: en_GB.UTF-8
  value of $XMODIFIERS: nil
  locale-coding-system: utf-8-unix
  default-enable-multibyte-characters: t

Major mode: Help

Minor modes in effect:
  tooltip-mode: t
  tool-bar-mode: t
  mouse-wheel-mode: t
  menu-bar-mode: t
  file-name-shadow-mode: t
  global-font-lock-mode: t
  font-lock-mode: t
  blink-cursor-mode: t
  global-auto-composition-mode: t
  auto-composition-mode: t
  auto-encryption-mode: t
  auto-compression-mode: t
  line-number-mode: t
  transient-mark-mode: t
  view-mode: t

Recent input:
<help-echo> <help-echo> <help-echo> <help-echo> <escape>
x C-g C-h h <help-echo> <next> <prior> <down> <down>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1> <mouse-1>
<mouse-1> <mouse-1> <mouse-1> <escape> : ( s e t q
SPC f o n t - l o g SPC n i l ) <return> <prior> <escape>
x f o n t - s h <tab> <return> <down-mouse-1> <mouse-movement>
<mouse-movement> <mouse-movement> <mouse-movement>
<mouse-movement> <drag-mouse-1> <down-mouse-1> <mouse-1>
<prior> <prior> <up> <up> <up> <up> <up> <up> <up>
<up> <up> <up> <up> <up> <up> <up> <up> <up> <up> C-@
<down> <down> <down> <down> <down> <down> <down> <down>
<down> <down> <down> <down> <down> <down> <down> <down>
<down> <down> <down> <down> <down> <down> <down> <down>
<down> <down> <down> <down> <down> <down> <down> <down>
<down> <down> <down> <down> <down> <down> <down> <down>
<down> <down> <down> <down> <down> <down> <down> <down>
<down> <down> <down> <down> <down> <down> <down> <down>
<down> <down> <down> <down> <down> <down> <escape>
w <help-echo> <help-echo> <help-echo> <help-echo> <escape>
x r e p o r t - e m a <tab> <return>

Recent messages:
For information about GNU Emacs and the GNU system, type C-h C-a.
Quit
Loading vc-cvs...done
View mode: type C-h for help, h for commands, q to quit.
nil
Mark activated
Saved text from "font for: (4121 . mule-unicode-0100-24ff"




[-- Attachment #3: Type: message/rfc822, Size: 1956 bytes --]

From: Chong Yidong <cyd@stupidchicken.com>
To: Juanma Barranquero <lekktu@gmail.com>
Cc: Jason Rumney <jasonr@gnu.org>, Kenichi Handa <handa@m17n.org>, 1990-done@emacsbugs.donarmstrong.com, 1952-done@emacsbugs.donarmstrong.com
Subject: Re: bug#1990: 23.0.60; Emacs trying too hard to find fonts that don't  exist
Date: Sat, 24 Jan 2009 17:44:13 -0500
Message-ID: <874ozowchu.fsf@cyd.mit.edu>

Juanma Barranquero <lekktu@gmail.com> writes:

>> That log message is emitted just before calling fontset_font (which is
>> where we would check whether fonts are known not to exist).
>
> OK.
>
> Then, please install your change. It's a big improvement.

I have installed a similar patch in CVS.