Re: [lmi] Better error messages for the GUI test failures

[Vadim Zeitlin]

On Thu, 26 Feb 2015 19:56:34 +0000 "Murphy, Kimberly" <address@hidden> wrote:

MK> Vadim Zeitlin wrote:
MK> 
MK> > On Thu, 12 Feb 2015 18:17:23 +0000 "Murphy, Kimberly" wrote:
MK> > MK> Vadim Zeitlin wrote:
MK> > MK>
MK> > MK> Using revision 6109 along with the updated wx trunk designated
MK> > MK> here:
MK> > MK>  http://lists.nongnu.org/archive/html/lmi/2015-01/msg00043.html
MK> > MK>
MK> > MK> the validate output tests no longer fail:
MK> > [...]
MK> > MK> The messages have improved. A number of improvements were made
MK> > MK> between revision 6086 and 6109. Can we pinpoint what the issue
MK> > MK> might have been?
MK> >
MK> >  Unfortunately I've never been able to reproduce it (which is a concern
MK> > on
MK> > its own) and looking at the commits between those two, nothing looks
MK> > obviously responsible. If you have the possibility to do it, it would be
MK> > great if you could please try bisecting the changes to find the exact
MK> > commit that fixed the problem.
MK> 
MK> Nothing jumped out at me either. Revision 6093 is the first commit
MK> where ‘validate_output_illustration’ runs successful:
MK>  http://svn.sv.gnu.org/viewvc/?view=rev&root=lmi&revision=6093

 Thanks for finding this! Unfortunately I still have no idea why/how did
this change to wx_test_pdf_create.cpp fixed the test from nominally
completely independent wx_test_validate_output.cpp file. The tests from the
latter file do run right after the tests from the former one, so I suspect
that the old version showed some unexpected dialog or otherwise changed the
program UI behaviour, but, again, I don't see any concrete mechanism by
which this could happen and am still unable to reproduce the bug. Does
anything unusual result from using "File|Print preview" while viewing an
illustration in lmi on your machine by chance? This is a complete shot in
the dark, but [not] running "File|Print preview" [any more] seems like the
only change in r6093 that could possibly affect anything outside of the
test it modified itself.

 In any case, I think it's not really productive to continue spending time
on trying to debug this, even though I dislike having unexplained things
happening in the program, I simply don't see what else could be done and,
after all, it does work now...

 Regards,
VZ