I see this issue every so often; particularly when non-Latin characters
are to be displayed.  For instance, a recent upgrade of libc6 crashed
when [apt-listchanges's pager] 'most' reached the line

/--------
|   * New Malayalam debconf translation, by Sajeev പിആര്‍.
\--------

in the changelog.  However, this happened only for one xterm (226-1 on
amd64), but not on another (225-1 on i386), both using the same X
server (xserver-xorg 1:7.2-5, on the i386).

Here's the error message:

/--------
| xterm:  warning, error event received:
| X Error of failed request:  BadValue (integer parameter out of range for 
operation)
|   Major opcode of failed request:  75 (X_PolyText16)
|   Value in failed request:  0x0
|   Serial number of failed request:  3780
|   Current serial number in output stream:  3783
\--------



Relevant environment:

/--------amd64 (fails)
| SHELL=/bin/bash
| TERM=xterm
| LANG=en_GB.UTF-8
| LANGUAGE=en_GB:en_US:en_GB:en
| DISPLAY=localhost:10.0
\--------

/--------i386 (succeeds)
| SHELL=/bin/bash
| TERM=xterm
| LANG=en_GB
| DISPLAY=:0.0
\--------

The two XTerms share their X resources, and both have UTF-8 checked and
disabled in the mouse-3 menu; TrueType fonts unchecked and disabled.

It looks like something is amiss in the Unicode/font handling somewhere,
perhaps?


Further investigation: changing LANG seems to make a difference, as
follows:

/--------
| $ LANG=en_GB.UTF-8 xterm -e zcat /usr/share/doc/libc6/changelog.Debian.gz 
| xterm:  warning, error event received:
| X Error of failed request:  BadValue (integer parameter out of range for 
operation)
|   Major opcode of failed request:  75 (X_PolyText16)
|   Value in failed request:  0x0
|   Serial number of failed request:  240
|   Current serial number in output stream:  2128
\--------

/--------
| $ LANG=en_GB xterm -e zcat /usr/share/doc/libc6/changelog.Debian.gz 
\--------
  (no output)


This result is consistent on both the machines I am using.

Does this help?


Another common case where I see the crash is when debconf uses whiptail,
though I'm finding this harder to write a reproducible test.

Reply via email to