Re: [XForms] fl_set_font_name() and segmentation fault error

[iprmaster]

Hi Jens,

sorry for the late answer, but I had a couple of days off. However, I
can say that the problem is not solved yet, otherwise I would have
written an email for the records :)

About your questions:

- wrapping my simple program within a valgrind call yields the X error
  but does not produce the segmentation fault any more (see
  valgrind.log), a quite common behaviour, as valgrind (according to my
  little knowledge) modifies the memory layout. This, in turn, can
  change the behaviour of pieces of code containing undefined
  behaviours. There should be some of them in either X or Xforms
  library, as valgrind's error messages like "Invalid read of size 8"
  shows :) 

- I am pretty sure that I am picking the right xforms version, although
  I can say that the "stable" version (1.0.90-r1) worked like a charm
  just right before I updated the X-server to the version 1.6.3

- I believe that the problem is related to functions which look for
  the font requested by the application. Indeed, if run the small
  application in within a ssh-tunnel from a PC with a different
  x-server (1.6.4 instead of 1.6.3), some valgrind errors disappear and
  only some "Conditional jump or move depends on uninitialised value"
  are still there. (see valgrind_remote.log)

- if I move the statements

  fl_set_icm_color(FL_COL1, 204, 204, 204);
  fl_set_icm_color(FL_TOP_BCOL, 230, 230, 230);
  fl_set_icm_color(FL_RED, 210, 0, 0);
  fl_set_icm_color(FL_GREEN, 0, 210, 0);
  fl_set_icm_color(FL_BLUE, 45, 45, 255);

  after the fl_initialize() call (as you do in your demos), only
  strncpy-related errors appears (see valgrind_remote2.log).

By the way, on which system are you running your tests? It may be we
are dealing with an X-error, which in turn may be due to
distribution-related patches.

Any additional hint will be welcome.

Best regards,
Max


On Sun, 25 Oct 2009 01:10:35 +0200
Jens Thoms Toerring <address@hidden> wrote:

> Hi Max,
> 
>     any new findings on your side? I unfortunately can't do
> much without a bit more of information - as I wrote the test
> program you sent works for me without any of the problems you
> encountered (segmentation fault and/or X error) so I've got
> nothing I could really use for debugging. I don't even know
> what you did do to get an X error instead of a segmentation
> fault (if you did), so I've no idea where to even start. Or
> is the problem already solved? In that case it might be use-
> ful for others if you would give a short description of what
> you found out.
>                            Best regards, Jens

valgrind.log
Description: Text Data

valgrind_remote.log
Description: Text Data

valgrind_remote2.log
Description: Text Data