bug#19496: 25.0.50; term.c, produce_composite_glyh, cmp->pixel_width uninitialized

bug#19496: 25.0.50; term.c, produce_composite_glyh, cmp->pixel_width uninitialized

[Stephen C. Gilardi]

emacs -Q
open an emacs lisp file that contains (lambda ... )
turn on prettify-symbols-mode
the lambda form is converted to (λ ... ), and should be displayed that way, but instead it's displayed as
(\
λ
 ...)

This commit is the first on master to show the problem: 20791069fa34b486c018ba7f27982bdc6ad2a4ea (found by git bisect)

The problem is the calculation of the pixel_width of the composition in the buffer. It's set to the pixel_width of the composition in the composition table, but that field is not initialized. (There is no mention of pixel_width anywhere in composite.c where the table entries are made.) When I check the value of cmp->pixel_width, it's somewhere near MAX_SHORT. It should be a small integer. Prior to this commit, in my setup, the value is 1 which avoids the problem, but isn't a credible pixel width.

An apparently very wide pixel_width for the composition explains the display anomaly. The composition appears to be thousands of characters wide and if that were true, the anomalous display would have been appropriate.

I don't see an obvious fix. There seems to be some confusing use in the code between "width in columns" and "width in pixels". It's possible that the pixel_width field in "struct it" is not named precisely for how it’s used--in a terminal, the width may actually be used as a number of columns rather than a number of pixels.


In GNU Emacs 25.0.50.1 (x86_64-unknown-linux-gnu)
 of 2015-01-03 on localhost
Repository revision: 20791069fa34b486c018ba7f27982bdc6ad2a4ea
System Description:	Ubuntu 14.04.1 LTS

Configured using:
 `configure --without-x'

Configured features:
SOUND GPM DBUS NOTIFY LIBSELINUX GNUTLS LIBXML2 ZLIB

Important settings:
  value of $LANG: en_US.UTF-8
  locale-coding-system: utf-8-unix

Major mode: Emacs-Lisp

Minor modes in effect:
  diff-auto-refine-mode: t
  tooltip-mode: t
  global-eldoc-mode: t
  electric-indent-mode: t
  prettify-symbols-mode: t
  menu-bar-mode: t
  file-name-shadow-mode: t
  global-font-lock-mode: t
  font-lock-mode: t
  auto-composition-mode: t
  auto-encryption-mode: t
  auto-compression-mode: t
  line-number-mode: t

Recent messages:
For information about GNU Emacs and the GNU system, type C-h C-a.
Prettify-Symbols mode enabled
user-error: Beginning of history; no preceding item

Load-path shadows:
None found.

Features:
(shadow sort gnus-util mail-extr emacsbug message dired format-spec
rfc822 mml mml-sec mm-decode mm-bodies mm-encode mail-parse rfc2231
mailabbrev gmm-utils mailheader sendmail rfc2047 rfc2045 ietf-drums
mm-util help-fns mail-prsvr mail-utils regexp-opt vc-git diff-mode
easymenu easy-mmode xterm time-date tooltip eldoc electric uniquify
ediff-hook vc-hooks lisp-float-type tabulated-list newcomment elisp-mode
lisp-mode prog-mode register page menu-bar rfn-eshadow timer select
mouse jit-lock font-lock syntax facemenu font-core frame cham georgian
utf-8-lang misc-lang vietnamese tibetan thai tai-viet lao korean
japanese hebrew greek romanian slovak czech european ethiopic indian
cyrillic chinese case-table epa-hook jka-cmpr-hook help simple abbrev
minibuffer nadvice loaddefs button faces cus-face macroexp files
text-properties overlay sha1 md5 base64 format env code-pages mule
custom widget hashtable-print-readable backquote make-network-process
dbusbind gfilenotify multi-tty emacs)

Memory information:
((conses 16 82152 5459)
 (symbols 48 17585 0)
 (miscs 40 36 102)
 (strings 32 12557 4834)
 (string-bytes 1 332483)
 (vectors 16 8665)
 (vector-slots 8 349036 23681)
 (floats 8 65 657)
 (intervals 56 259 25)
 (buffers 976 12)
 (heap 1024 37582 1315))

bug#19496: 25.0.50; term.c, produce_composite_glyh, cmp->pixel_width uninitialized

[Eli Zaretskii]

> From: "Stephen C. Gilardi" <scgilardi@gmail.com>
> Date: Sat, 3 Jan 2015 09:20:52 -0500
> 
> emacs -Q
> open an emacs lisp file that contains (lambda ... )
> turn on prettify-symbols-mode
> the lambda form is converted to (λ ... ), and should be displayed that way, but instead it's displayed as
> (\
> λ
>  ...)
> 
> This commit is the first on master to show the problem: 20791069fa34b486c018ba7f27982bdc6ad2a4ea (found by git bisect)

I think that commit was mistaken, and should be reverted, because
before it this worked correctly.  CC'ing Paul, who made the change,
and Handa-san, who probably wrote it.

Paul, was there a use case where you saw a problem with the original
code?  If so, can you show it?

> I don't see an obvious fix.

Why do you think the previous code is not TRT?

> There seems to be some confusing use in the code between "width in columns" and "width in pixels". It's possible that the pixel_width field in "struct it" is not named precisely for how it’s used--in a terminal, the width may actually be used as a number of columns rather than a number of pixels.

On a text terminal, each glyph is 1 pixel wide.  So on a text terminal
these two are identical.

bug#19496: 25.0.50; term.c, produce_composite_glyh, cmp->pixel_width uninitialized

[Paul Eggert]

Eli Zaretskii wrote:
> Paul, was there a use case where you saw a problem with the original
> code?  If so, can you show it?

Sorry, that was a thinko on my part.  I reverted the change in master as commit 
5395106.  Thanks for reporting it.