Re: [Denemo-devel] Testsuite, step 1

[Richard Shann]

On Thu, 2013-11-14 at 17:23 +0100, Éloi Rivard wrote:
> About the loading of the last file, I look at this.
> 
> About the thumbnail, how can I reproduce this bug ? I did not see such
> a message.
I think you must have the Denemo.prefs->enable_thumbnails set false.
I see why it is failing - contrary to what I thought the invocation of
denemo for thumbnails is via this line in printview.c

g_build_filename (get_system_bin_dir (), "denemo", NULL)

I guess that this is invoking the same executable as it is called from,
and hence it will fail because it tries to open the file in
non-interactive mode. At one time get_bin_dir() used to
find /usr/local/bin which is why I thought the crash was coming from the
earlier version of denemo, which would have been puzzling.

So, top priority is to get file load working in non-interactive mode.
The turning off of the signal handlers should not be done in the case
where there is no gui. It happens in close_gui() and will be done in
file.c when a file is going to be loaded (because it can't be drawn
while its data structures are being created IIRC).

HTH

Richard


> 
> About loading a file in non-interactive mode. I know it fails, this is
> because the file loading depends on the gui, that is not launched.
> This is the first thing I want to fix after I set up the test suite
> (this would be the second test).
> 
> 
> 
> 2013/11/14 Richard Shann <address@hidden>
>         On Thu, 2013-11-14 at 13:47 +0000, Richard Shann wrote:
>         > On Thu, 2013-11-14 at 13:21 +0100, Éloi Rivard wrote:
>         > > I just made some tests on the command-hash-table and the
>         master branch
>         > > at the same time. Search seems to have the same behavior.
>         > > Is it ok for you if I merge it in master ?
>         >
>         > Yes, it seems ok.
>         
>         
>         Of course, I spoke too soon. The automatic loading of the last
>         file
>         doesn't appear to be happening any more.
>         And the thumbnailer (which runs on *exit* not entry) is
>         failing - note I
>         am running an earlier version of Denemo as the thumbnailer
>         (the one
>         installed on my system), so it seems strange this could start
>         failing.
>         
>         The message from (I think) the thumbnailer instance of Denemo
>         is this:
>         
>         Denemo 1_1_1
>         Loaded
>         keymap /home/rshann/.denemo-1.1.1/actions/Default.commands
>         
>         (denemo:6825): GLib-GObject-WARNING **: invalid (NULL) pointer
>         instance
>         
>         (denemo:6825): GLib-GObject-CRITICAL **:
>         g_signal_handlers_block_matched: assertion
>         `G_TYPE_CHECK_INSTANCE
>         (instance)' failed
>         
>         the signal handler block is (I suspect) the one done to
>         prevent drawing
>         while a file is being loaded.
>         
>         Here are two runs with non-interactive set, loading the file
>         junk.denemo
>         and executing (d-Quit), the first is the earlier version of
>         denemo
>         1.1.1, while the one in ~/local/bin is the latest from git:
>         
>         denemo junk.denemo -n -a "(d-Quit)"
>         
>         xmlsource == /home/rshann/.denemo-1.1.1/denemorc
>         audio driver is 'portaudio' 0
>         Initializing Fluidsynth
>         fluidsynth: warning: Failed to pin the sample data to RAM;
>         swapping is possible.
>         
>         ** (denemo:6890): WARNING **: The default fluidsynth soundfont
>         has been loaded
>         Initializing PortAudio backend
>         ALSA lib pcm.c:2217:(snd_pcm_open_noupdate) Unknown PCM
>         cards.pcm.rear
>         ALSA lib pcm.c:2217:(snd_pcm_open_noupdate) Unknown PCM
>         cards.pcm.center_lfe
>         ALSA lib pcm.c:2217:(snd_pcm_open_noupdate) Unknown PCM
>         cards.pcm.side
>         ALSA lib pcm_dmix.c:957:(snd_pcm_dmix_open) The dmix plugin
>         supports only playback stream
>         opening output device 'ALSA: HDA Intel PCH: VT2020 Analog
>         (hw:0,0)'
>         MIDI driver is 'portmidi'
>         initializing PortMidi backend
>         opening input device 'ALSA: USB Keystation 49e MIDI 1'
>         opening output device 'ALSA: Midi Through Port-0'
>         Version 1_1_1Arranger Profile
>         stopping playback
>         Removed 0
>         that's it, i quit!
>         destroying PortAudio backend
>         destroying PortMidi backend
>         address@hidden:~$ ./local/bin/denemo junk.denemo -n -a
>         "(d-Quit)"
>         
>         xmlsource == /home/rshann/.denemo-1.1.1/denemorc
>         Denemo 1_1_1
>         Loaded
>         keymap /home/rshann/.denemo-1.1.1/actions/Default.commands
>         
>         (denemo:6900): GLib-GObject-WARNING **: invalid (NULL) pointer
>         instance
>         
>         (denemo:6900): GLib-GObject-CRITICAL **:
>         g_signal_handlers_block_matched: assertion
>         `G_TYPE_CHECK_INSTANCE (instance)' failed
>         Segmentation fault
>         
>         This appears to be the invocation itself rather than the
>         thumbnailer
>         that has crashed. I think there is an option to turn off
>         thumbnailing,
>         which would be good while trying to test.
>         
>         Richard
>         
>         
>         >
>         > Richard
>         >
>         >
>         > >
>         > > 2013/11/14 Richard Shann <address@hidden>
>         > >         On Thu, 2013-11-14 at 12:56 +0100, Éloi Rivard
>         wrote:
>         > >         >
>         > >         > Hum, I am not sure I understand the behavior
>         search should
>         > >         have.
>         > >         >
>         > >         > When checked, it should search in tooltips, when
>         not it
>         > >         should search
>         > >         > on the name ?
>         > >
>         > >         At first I think I had it working on searching the
>         label, then
>         > >         I
>         > >         alphabeticalized the list by label so I switched
>         to making the
>         > >         search
>         > >         use the internal command name (as no search was
>         needed for the
>         > >         label).
>         > >         Then I discovered that alphabeticalizing the list
>         was still
>         > >         causing
>         > >         troubles (with user-creation of commands) so I
>         turned that
>         > >         off. It may
>         > >         be since then that it is behaving strangely, or it
>         may be
>         > >         since you
>         > >         re-organized the gtk-list/g_hash thing. Sorry to
>         be so vague.
>         > >         I didn't
>         > >         test it for a while, then one time I needed it ...
>         > >
>         > >         Richard
>         > >
>         > >
>         > >
>         > >
>         > >
>         > > --
>         > > Éloi Rivard - address@hidden
>         > >
>         > > « On perd plus à être indécis qu'à se tromper. »
>         > >
>         >
>         >
>         >
>         
>         > _______________________________________________
>         > Denemo-devel mailing list
>         > address@hidden
>         > https://lists.gnu.org/mailman/listinfo/denemo-devel
>         
>         
>         
> 
> 
> 
> -- 
> Éloi Rivard - address@hidden
>         
> « On perd plus à être indécis qu'à se tromper. »
>