Re: [Denemo-devel] Denemo bugs

[Richard Shann]

On Fri, 2014-10-17 at 13:42 +0100, Richard Shann wrote:
> It would be best to remove that version - 

I have now used Denemo for about half and hour now without further
crashes, so I think my fixes to the header file inclusion are good. In
that case, the bug only happens on 64-bit builds, so the darwin build
you posted would have been ok after all (sorry!).

And indeed Bert Wiest did not report crashes with that version.
You mentioned installing hacks to avoid threading problems - I guess
that would be the mutex unlocks? Are the threading problems just on
exiting Denemo or do you see problems while using Denemo?

I fairly often get Denemo hanging on exit in a pthread routine on my
Debian system, I haven't seen this happen on the windows binaries, but
then I don't get very long trying them. (People don't report this as a
problem on Windows, which is stronger evidence). I did change the way
Denemo exits for the G_OS_WIN32 case, skipping the audio shutdown and
calling _exit() directly, I think.

Richard




> the crashes are very
> unpredictable, Eloi has done a massive re-organization of the code.
> Richard
> On Fri, 2014-10-17 at 07:19 -0500, Jeremiah Benham wrote:
> > Oh no. This was a git snapshot made around 4pm CST. So this may have
> > the bug you mentioned in another thread.
> > 
> > Jeremiah
> > 
> > On Oct 17, 2014 3:09 AM, "Richard Shann" <address@hidden>
> > wrote:
> >         On Thu, 2014-10-16 at 20:36 -0500, Jeremiah Benham wrote:
> >         > Can you please test the following:
> >         >
> >         >
> >         
> > http://denemo.org/~jjbenham/gub/uploads/denemo-0.0.0-0.darwin-x86.tar.bz2
> >         >
> >         > I have reverted back to gtk2 so that the fonts show in the
> >         pallets.
> >         
> >         Just to check - is this built from 1.1.8 (or at least a
> >         version prior to
> >         12th Oct when master became insidiously broken?
> >         
> >         Richard
> >         
> >         
>