* Sven Joachim <svenj...@gmx.de> [2018-12-29 19:20]:

> On 2018-12-21 20:54 +0100, Alexander Meyer wrote:
> 
>> * Sven Joachim <svenj...@gmx.de> [2018-12-21 20:38]:
>> 
>>> On 2018-12-21 20:16 +0100, Alexander Meyer wrote:
>>> 
>>>> I've found a bug report from Arch Linux that looks similar:
>>>> https://bugs.archlinux.org/task/61115

[...]

>>> I could also still reproduce it, that's why I did not close #916349 in
>>> the xterm 340-1 upload.
>> 
>> I couldn't tell whether that bug was the same as mine, as none of the
>> error messages mentioned there came up for me, also that bug doesn't
>> talk about segfaulting.
> 
> Oh, indeed.  But neither does the Archlinux bug report, it only mentions
> the dreadful "X Error of failed request" which aborts xterm.

That's right, indeed. Maybe mentioning the Arch Linux report caused
confusion rather than it helped. I'm sorry.

> And I could reproduce that with xterm 338 trough 340, but not your
> segfault.
> 
>> If you think this is the same bug, then feel free to merge, of course.
> 
> If you still get a segfault in xterm 341 (uploaded today), please tell
> me and I will unmerge and reopen your bug.

Unfortunately, I still get the same segfault with 341. But I've just
noticed that the behaviour changes depending on whether I have a
fonts.conf file enabled or not.

I normally use the fonts.conf file from
https://gist.github.com/dcrystalj/d1c1ceacf0d6fc9a0556
installed in ~/.config/fontconfig/fonts.conf

This is the behaviour I get across xterm versions:
(everything with libfontconfig1 2.13.1-2)

fonts.conf enabled:
337: works
338: segfault
340: segfault
341: segfault

fonts.conf disabled:
337: works
338: aborts with error message "BadLength (poly request too large or
     internal Xlib length error)" as described in bug 916349
340: works! (as opposed to what is reported in bug 916349)
341: works


So the segfault only happens with that fonts.conf. To be honest, I know
next to nothing about fontconfig or font handling in X. I've installed
that file some years ago without knowing what it actually does because
it made some fonts look nicer in Firefox etc. and happened to have no
ill side effects until now. Maybe I could even get along without it, but
that wouldn't resolve this bug, of course.

Best
Alex

Reply via email to