|
| From: | Richard Shann |
| Subject: | Re: [Denemo-devel] New Binaries? |
| Date: | Fri, 29 May 2015 18:10:40 +0100 |
I think I have found a description of our problem:
https://mail.gnome.org/archives/gtk-list/2012-July/msg00000.html
this guy even suggests a fix:
8><8><8><8><8><8><8><8><8><8><8><8><8><8><8><8><8><8><8><8><8><8><8><8><8><8><
diff --git a/gdk/win32/gdkevents-win32.c b/gdk/win32/gdkevents-win32.c
index 2bc7c12..3505fe2 100644
--- a/gdk/win32/gdkevents-win32.c
+++ b/gdk/win32/gdkevents-win32.c
@@ -3369,9 +3369,10 @@ gdk_event_dispatch (GSource *source,
{
GdkEvent *event;
+ _gdk_win32_display_queue_events (_gdk_display);
+
GDK_THREADS_ENTER ();
- _gdk_win32_display_queue_events (_gdk_display);
event = _gdk_event_unqueue (_gdk_display);
if (event)
8><8><8><8><8><8><8><8><8><8><8><8><8><8><8><8><8><8><8><8><8><8><8><8><8><8><
which is moving the call to _gdk_win32_display_queue_events() outside of
the GDK_THREADS_ENTER (); section.
The thread ends with Emanuele Bassi sending him off to the jungle of
gnome bug trackers ...
This report is nearly a year old. Can it really be broken so long?
Richard
On Fri, 2015-05-29 at 09:34 -0500, Jeremiah Benham wrote:
> I have built another one based off the gtk win32 bundle but it also
> crashed when moving the window. I don't know what else to do here.
> Once you mentioned a native windows tool that was similar to mxe.
> Would it be easier to natively compile it?
>
> Jeremiah
>
> On May 25, 2015 12:07 PM, "Richard Shann" <address@hidden>
> wrote:
> On Sun, 2015-05-24 at 23:54 -0500, Jeremiah Benham wrote:
> > I have found a page on gtk's website that has all the
> version numbers
> > they recommend for gtk3 and win32.
> That sounds promising, let me know when there is something to
> test.
> > I am going to use this as a template and rebuild. I did
> retest and I
> > believe it worked in vista but not in XP.
>
> Hmm, I was able to drag the main display a short distance
> before it
> crashed on the last version, instead of it crashing straight
> away.
>
> Richard
>
> >
> > Jeremiah
> >
> > On May 24, 2015 10:41 AM, "Richard Shann"
> <address@hidden>
> > wrote:
> > I saw that you have tried another build dated 25th
> May, so I
> > tested this
> > one too from the zip version and it still will not
> allow any
> > window
> > re-size nor run LilyPond. From all I have read it
> would seem
> > likely that
> > some combination of library versions and compiler
> versions
> > will work,
> > but I don't know how to find out which. Have you
> been able to
> > re-test
> > the one that appeared to work on windows vista?
> >
> > Richard
> >
> >
> > On Fri, 2015-05-22 at 08:47 -0500, Jeremiah Benham
> wrote:
> > >
> > >
> > > On Fri, May 22, 2015 at 4:12 AM, Richard Shann
> > > <address@hidden> wrote:
> > > On Thu, 2015-05-21 at 15:26 -0500,
> Jeremiah Benham
> > wrote:
> > > > I created new mingw binaries. This time
> it is
> > based on gtk3.
> > >
> > > I've downloaded and tested
> > >
> >
> http://denemo.org/~jjbenham/gub/uploads/denemo-mingw-0.0.0.zip
> > >
> > > running Denemo.bat it complains that
> libaubio-4.dll
> > is not
> > > found - I
> > > moved it from /lib to /bin and the program
> then
> > started
> > > correctly.
> > >
> > >
> > >
> > > Ok. Thanks for letting me know.
> > >
> > >
> > > The good news is that the fonts are now
> found, and
> > the nice
> > > gtk3
> > > features are present.
> > > However, LilyPond crashes while processing
> a file
> > and Denemo
> > > crashes the
> > > moment you try to re-size a window, both
> symptoms
> > exactly as
> > > before.
> > >
> > > The crash when trying to move or re-size a
> window
> > has probably
> > > not
> > > changed because, despite moving to gtk3 I
> guess that
> > the
> > > version of glib
> > > is unaltered?
> > >
> > >
> > >
> > > I used 2.43 for both. I will try downgrading and
> rebuilding.
> > >
> > >
> > > Whether this is related to lilypond not
> working I'm
> > not sure
> > >
> > >
> > >
> > > I will upgrade it to 2.18.2 or whatever the latest
> stable of
> > lilypond.
> > >
> > >
> > > - the
> > > attached screenshot on the blue background
> is from
> > running gdb
> > > lilypond.exe on a file called fbtest.ly
> which
> > succeeds in
> > > earlier
> > > versions of the denemo installer. The
> crash in this
> > case is
> > > reported in
> > > a memory allocation routine.
> > >
> > > So the most hopeful line might be to look
> at which
> > version of
> > > glib is
> > > involved in the last two builds. I guess
> there are
> > also lower
> > > level
> > > things (people talk about pthreads and the
> > compiler's
> > > libraries ...)
> > > which could be giving us the trouble.
> > >
> > > We could install a piece of code at the
> start of
> > main() in
> > > Denemo that
> > > - never executes the rest of Denemo's
> > > code. Then trying to resize the window
> should
> > confirm that the
> > > bug is
> > > not connected with Denemo's use of
> threads.
> > >
> > >
> > > Could also you try to resize the dialog window
> that asks you
> > if you
> > > want to import your old prefs?
> > >
> > >
> > > I will rebuild now.
> > >
> > >
> > > Jeremiah
> > >
> > >
> > >
> > >
> > >
> > > Richard
> > >
> > >
> > >
> > > >
> > > > Jeremiah
> > > >
> > > > On May 13, 2015 6:51 AM, "Richard Shann"
> > > <address@hidden>
> > > > wrote:
> > > > On Tue, 2015-05-12 at 15:40
> -0500,
> > Jeremiah Benham
> > > wrote:
> > > > > What do you think is causing
> the missing
> > don't
> > > >
> > > > I've done some gdb debugging on
> the fonts
> > thing: I
> > > noticed
> > > > that one of
> > > > the buttons in the horizontal
> palette, the
> > one for
> > > entering
> > > > notes in
> > > > chords *was* showing the two
> quarter notes
> > in the
> > > correct
> > > > font. Looking
> > > > at its label I see that it uses
> the gtk
> > markup
> > > language to set
> > > > the font
> > > > explicitly, where the other
> palette
> > buttons just
> > > assume the
> > > > font has
> > > > been set. The menu items and
> tooltips will
> > also just
> > > assume
> > > > the font is
> > > > set I think. The call to set the
> default
> > font is
> > > towards the
> > > > end of
> > > > view.c
> > > > gtk_settings_set_string_property
> > (settings,
> > > "gtk-font-name",
> > > > Denemo.prefs.fontspec->str,
> "denemo");
> > > >
> > > > this appears to be called twice
> when
> > starting the
> > > file, but
> > > > the
> > > > backtrace is obscure. I tried
> calling it
> > again from
> > > within gdb
> > > > but this
> > > > gives a crash during the call,
> perhaps
> > because of
> > > not being in
> > > > the right
> > > > part of the gtk_main loop. In
> fact, mostly
> > this gdb
> > > > environment is
> > > > obscure - the symbol Denemo
> (which should
> > be the
> > > root object
> > > > of all the
> > > > Denemo data) prints as 0, and
> when trying
> > to access
> > > > Denemo.project
> > > > (which should be the current
> score) gdb
> > says Denemo
> > > is not a
> > > > structure.
> > > > The info threads just lists a
> bunch of
> > threads
> > > in .dlls
> > > > (whereas our
> > > > threads now have proper names).
> > > >
> > > > So this fonts problem is not to
> do with
> > failing to
> > > install
> > > > fonts, but
> > > > failing to make gtk use the
> correct font
> > as the
> > > default.
> > > >
> > > > > and the crashing?
> > > >
> > > > this happens the instant you try
> to
> > re-size a
> > > window. I found
> > > > out
> > > > something else too - LilyPond is
> crashing,
> > windows
> > > throws up a
> > > > window to
> > > > say it has stopped working,
> while denemo
> > itself
> > > continues to
> > > > work ok.
> > > > I've run it from the command
> line and it
> > gets as far
> > > as
> > > > "Preprocessing
> > > > graphical objects" and then
> complains
> > about
> > > std::bad_alloc()
> > > >
> > > > Perhaps we might have better
> luck with
> > gtk3 - at
> > > least they
> > > > are actively
> > > > working on that.
> > > > >
> > > > > I am having trouble creating
> binaries
> > for Darwin.
> > > I did
> > > > natively
> > > > > compile denemo on Darwin but
> the
> > denemo.ttf don't
> > > is not
> > > > displaying on
> > > > > the side or in the menu.
> Everything else
> > seems
> > > fine. This
> > > > was compiled
> > > > > against gtk3. I don't know if
> the
> > problem will be
> > > gone if I
> > > > use gtk2.
> > > > > What do you recommend I do?
> > > >
> > > > This seems to be the failure to
> get gtk to
> > take
> > > notice of the
> > > > desired
> > > > default font that I've described
> above.
> > Apparently
> > > there
> > > > are .ini type
> > > > files in which the desired font
> can be
> > specified ...
> > > >
> > >
> >
> https://developer.gnome.org/gtk3/3.8/GtkSettings.html
> > > >
> > > >
> > > > Richard
> > > >
> > > >
> > > >
> > > > >
> > > > > Jeremiah
> > > > >
> > > > > On May 11, 2015 9:44 AM,
> "Richard Shann"
> > > > <address@hidden>
> > > > > wrote:
> > > > > I thought it might be
> good to
> > check the
> > > installer
> > > > version too
> > > > > - that
> > > > > behaves similarly, no
> fonts and
> > it works
> > > until you
> > > > try a
> > > > > window re-size,
> > > > > which is instant
> death.
> > > > >
> > > > > Richard
> > > > >
> > > > >
> > > > > On Wed, 2015-05-06 at
> 11:15
> > +0100, Richard
> > > Shann
> > > > wrote:
> > > > > > On Tue, 2015-05-05
> at 16:29
> > -0500,
> > > Jeremiah Benham
> > > > wrote:
> > > > > > > I upgraded gtk.
> Maybe this
> > will fix
> > > the problem.
> > > > I created
> > > > > new mingw
> > > > > > > binaries.
> > > > > > >
> > > > > > I've tested the zip
> file
> > version - it
> > > starts up
> > > > with the
> > > > > denemo.ttf font
> > > > > > not found, but notes
> can be
> > entered. Any
> > > attempt
> > > > at
> > > > > re-sizing a window
> > > > > > etc causes a crash.
> > Typesetting appears
> > > to cause a
> > > > crash in
> > > > > the lilypond
> > > > > > process started
> (Denemo
> > carries on
> > > working, but a
> > > > microsoft
> > > > > visual c++
> > > > > > runtime library
> window pops up
> > to warn
> > > of a crash,
> > > > but it
> > > > > never gets
> > > > > > filled in - see
> attachment).
> > > > > >
> > > > > > Richard
> > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > > > > Jeremiah
> > > > > > >
> > > > > > > On May 5, 2015
> 9:00 AM,
> > "Richard
> > > Shann"
> > > > >
> <address@hidden>
> > > > > > > wrote:
> > > > > > > On Tue,
> 2015-05-05
> > at 06:21
> > > -0500,
> > > > Jeremiah Benham
> > > > > wrote:
> > > > > > > > I have
> created the
> > mingw
> > > binaries. I
> > > > will work
> > > > > on the others
> > > > > > > today.
> > > > > > >
> > > > > > > I've
> downloaded and
> > run the
> > > zip file
> > > > version -
> > > > > unfortunately
> > > > > > > it crashes
> > > > > > > on
> starting - it
> > hangs the
> > > machine for
> > > > quite a
> > > > > while as it
> > > > > > > does this.
> I
> > > > > > > ran it
> under gdb and
> > saw that
> > > it seemed
> > > > to be some
> > > > > sort of
> > > > > > > thread
> > > > > > > problem -
> I turned
> > off the
> > > continuous
> > > > typesetting
> > > > > (by editing
> > > > > > > the denemo
> > > > > > > rc file)
> and it
> > started ok. I
> > > could
> > > > enter a few
> > > > > notes and they
> > > > > > > sounded
> > > > > > > on entry.
> When I
> > tried to
> > > typeset in the
> > > > Print
> > > > > View window it
> > > > > > > completely
> > > > > > > hung the
> machine,
> > requiring a
> > > hard
> > > > re-boot. This
> > > > > is on the
> > > > > > > Windows
> Vista
> > > > > > > laptop
> that runs
> > 1.2.2 without
> > > a
> > > > problem.
> > > > > > > I did
> another run,
> > this time
> > > without
> > > > trying
> > > > > typesetting -
> > > > > > > after a
> while
> > > > > > > it froze
> the
> > machine. I found
> > > that
> > > > ctrl-alt-del
> > > > > gave me access
> > > > > > > to the
> > > > > > > machine
> again and I
> > was able
> > > to take a
> > > > snapshot of
> > > > > the gdb
> > > > > > > console
> which
> > > > > > > I attach.
> It looks
> > like this
> > > is
> > > > threading problems
> > > > > inside the
> > > > > > > gtk system
> > > > > > > rather
> than some
> > denemo thread
> > > problem
> > > > (I was just
> > > > > changing
> > > > > > > between
> > > > > > > windows in
> Denemo
> > when it
> > > crashed,
> > > > rather than
> > > > > initiating
> > > > > > > playback,
> > > > > > > entering
> notes etc -
> > that's no
> > > > guarantee, of
> > > > > course, but my
> > > > > > > gut
> > > > > > > feeling).
> > > > > > > I think it
> would
> > help if we
> > > could get
> > > > rid of the
> > > > > code that
> > > > > > > prompts
> "Type
> > > > > > > ENTER..."
> - I know
> > you have
> > > tried
> > > > before, but
> > > > > evidently it is
> > > > > > > still
> > > > > > > there.
> > > > > > >
> > > > > > > Richard
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > > >
> > > > > > > > Jeremiah
> > > > > > > >
> > > > > > > > On May
> 1, 2015
> > 10:25 AM,
> > > "Richard
> > > > Shann"
> > > > > > >
> > <address@hidden>
> > > > > > > > wrote:
> > > > > > > >
> On Fri,
> > 2015-05-01
> > > at 09:42
> > > > -0500,
> > > > > Jeremiah Benham
> > > > > > > wrote:
> > > > > > > >
> > I now
> > have a gub
> > > branch with
> > > > upgraded
> > > > > glib but I
> > > > > > > am stuck
> on
> > > > > > > >
> nsis . I
> > > > > > > >
> > could
> > create a zip
> > > version
> > > > now.
> > > > > > > >
> > > > > > > >
> That would
> > be
> > > excellent -
> > > > people can try
> > > > > it out
> > > > > > > without
> > > > > > > >
> committing
> > > > > > > >
> > themselves.
> > > > > > > >
> > > > > > > >
> Is the
> > nsis problem
> > > related to
> > > > the new
> > > > > glib?
> > > > > > > >
> > > > > > > >
> Richard
> > > > > > > >
> > > > > > > >
> >
> > > > > > > >
> > Jeremiah
> > > > > > > >
> >
> > > > > > > >
> > On Apr
> > 29, 2015
> > > 8:20 AM,
> > > > "Jeremiah
> > > > > Benham"
> > > > > > > >
> >
> > > <address@hidden>
> > > > wrote:
> > > > > > > >
> >
> > I will
> > > work on it.
> > > > It may take
> > > > > a little
> > > > > > > longer
> than
> > > > > > > >
> I
> > > > > > > >
> >
> > expected.
> > > > > > > >
> >
> > > > > > > >
> >
> > Jeremiah
> > > > > > > >
> >
> > > > > > > >
> >
> > On Apr 29,
> > > 2015 7:59
> > > > AM,
> > > > > "Richard Shann"
> > > > > > > >
> >
> > > > <address@hidden>
> > > > > wrote:
> > > > > > > >
> >
> > > This seems
> > > > to be the
> > > > > latest
> > > > > > > version of
> glib:
> > > > > > > >
> >
> > > > > > > >
> >
> > > > > > > >
> > > > > > >
> > > > >
> > > >
> > >
> >
> https://developer.gnome.org/glib/glib-html-2.44.0.tar.gz
> > > > > > > >
> >
> > > > > > > >
> >
> > If
> > > that will
> > > > work it
> > > > > would be our
> > > > > > > best bet
> > > > > > > >
> for bug
> > > > > > > >
> >
> > > fixes.
> > > > > > > >
> >
> > > > > > > >
> >
> > > Richard
> > > > > > > >
> >
> > > > > > > >
> >
> > On
> > > Wed,
> > > > 2015-04-29 at
> > > > > 00:02 -0500,
> > > > > > > Jeremiah
> > > > > > > >
> Benham
> > > > > > > >
> >
> > > wrote:
> > > > > > > >
> >
> > >
> > > I'm
> > > > getting this
> > > > > error when
> > > > > > > trying to
> > > > > > > >
> compile
> > for
> > > > > > > >
> >
> > >
> > > mingw:
> > > > > > > >
> >
> > >
> > > building
> > > > package:
> > > > > mingw::denemo
> > > > > > > >
> >
> > >
> > > *** Stage:
> > > > download
> > > > > (denemo,
> > > > > > > mingw)
> > > > > > > >
> >
> > >
> > > *** Stage:
> > > > untar
> > > > > (denemo, mingw)
> > > > > > > >
> >
> > >
> > > *** Stage:
> > > > patch
> > > > > (denemo, mingw)
> > > > > > > >
> >
> > >
> > > *** Stage:
> > > > > autoupdate (denemo,
> > > > > > > mingw)
> > > > > > > >
> >
> > >
> > > *** Stage:
> > > > configure
> > > > > (denemo,
> > > > > > > mingw)
> > > > > > > >
> >
> > >
> > > *** Stage:
> > > > compile
> > > > > (denemo,
> > > > > > > mingw)
> > > > > > > >
> >
> > >
> > > Command
> > > > barfed:
> > > > > > > >
> >
> > >
> > > > > > > >
> > > > > > >
> > > > >
> > > >
> > >
> >
> cd
> /home/jjbenham/public_html/gub/target/mingw/build/denemo-git.savannah.gnu.org--denemo.git-master
> && make -j16
> > > > > > > >
> >
> > >
> > > Tail of
> > > > > > >
> > target/mingw/log/denemo.log
> > > > > > > >
> >>>>>>>>
> > > > > > > >
> >
> > >
> > > > make[1]: ***
> > > > > [all-recursive]
> > > > > > > Error 1
> > > > > > > >
> >
> > >
> > > > make[1]: Leaving
> > > > > directory
> > > > > > > >
> >
> > >
> > > > > > > >
> > > > > > >
> > > > >
> > > >
> > >
> >
>
> `/home/jjbenham/public_html/gub/target/mingw/build/denemo-git.savannah.gnu.org--denemo.git-master'
> > > > > > > >
> >
> > >
> > > > > > > >
> > > > > > >
> > > > >
> > > >
> > >
> >
>
> libaudiobackend.a(libaudiobackend_a-audiointerface.o):audiointerface.c:(.text+0x445):
> undefined reference to `_g_thread_try_new'
>
> libaudiobackend.a(libaudiobackend_a-audiointerface.o):audiointerface.c:(.text+0x8c5):
> undefined reference to `_g_cond_wait_until'
>
> libaudiobackend.a(libaudiobackend_a-dummybackend.o):dummybackend.c:(.text+0x68):
> undefined reference to
> `_g_cond_wait_untilibaudiobackend.a(libaudiobackend_a-audiointerface.o):audiointerface.c:(.text+0x445):
> undefined reference to `_g_thread_try_new'
> @
> > > > > > > >
> >
> > >
> > > > > > > >
> >
> > >
> > > On Apr 28,
> > > > 2015
> > > > > 12:34 PM,
> > > > > > > "Jeremiah
> > > > > > > >
> Benham"
> > > > > > > >
> >
> > >
> > > > >
> <address@hidden>
> > > > > > > wrote:
> > > > > > > >
> >
> > > > I
> > > > will do
> > > > > that tonight.
> > > > > > > >
> >
> > >
> > > > > > > >
> >
> > >
> > > > Jeremiah
> > > > > > > >
> >
> > >
> > > > > > > >
> >
> > > > On
> > > > Apr 28,
> > > > > 2015 7:06 AM,
> > > > > > > "Richard
> > > > > > > >
> Shann"
> > > > > > > >
> >
> > >
> > > > > > >
> > <address@hidden>
> > > wrote:
> > > > > > > >
> >
> > >
> > > > > Jeremiah,
> > > > > > > >
> >
> > >
> > > > > > > >
> >
> > >
> > > > The
> > > > > fixes I've
> > > > > > > made for
> > > > > > > >
> threading
> > > > > > > >
> >
> > >
> > > > > issues seem to
> > > > > > > be
> > > > > > > >
> effective
> > - and I
> > > > > > > >
> >
> > >
> > > > have
> > > > > fixed a
> > > > > > > critical
> bug
> > > > > > > >
> in copy &
> > > > > > > >
> >
> > >
> > > > > paste - I think
> > > > > > > we should
> > > > > > > >
> try for a
> > > > > > > >
> >
> > >
> > > > > early release of
> > > > > > > 1.2.4.
> > > > > > > >
> Are you
> > able
> > > > > > > >
> >
> > >
> > > > to
> > > > > generate
> > > > > > > binaries
> at
> > > > > > > >
> the
> > moment?
> > > > > > > >
> >
> > >
> > > > > > > >
> >
> > >
> > > > > Richard
> > > > > > > >
> >
> > >
> > > > > > > >
> >
> > >
> > > > > > > >
> >
> > > > > > > >
> >
> > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > > >
> > >
> _______________________________________________
> > > > > > Denemo-devel mailing
> list
> > > > > > address@hidden
> > > > > >
> > > >
> > https://lists.gnu.org/mailman/listinfo/denemo-devel
> > > > >
> > > > >
> > > >
> > > >
> > >
> > >
> > >
> > >
> >
>
>
| [Prev in Thread] | Current Thread | [Next in Thread] |