Re: Denemo

[Richard Shann]

On Sat, 2020-06-27 at 11:55 -0400, Eddy wrote:
> I managed to get this to run, although I'm not entirely sure why what
> I
> did works. Just a reminder that I'm running on 10.13.6 and satisfying
> dependencies with macports, but that the default "mkdir bin && cd
> bin;
> ../configure; make" gives a binary that segfaults.
> 
> If I use "./configure --enable-debug" in this process instead of just
> "./configure" then I can get it to run.

Looking at configure.ac I see this:

AC_ARG_ENABLE(
  debug,
  AS_HELP_STRING([--enable-debug], [enable debug code
@<:@default=no@:>@]),
  [
    if test "x$enableval" != "xno"; then
      build_debug=yes
      CFLAGS="-O0 -DDEBUG -g -Wall -Wextra -Wno-deprecated-declarations 
-Wno-sign-compare"
      CXXFLAGS="-O0 -DDEBUG -g -Wall -Wextra -Wno-deprecated-
declarations -Wno-sign-compare"
    fi
  ])

which I take to mean that --enable-debug turns the compiler
optimizations off (i.e. -O0).

To check if it is the compiler optimizer that is causing the crash (as
opposed to the setting of the -DDEBUG flag, which does introduce new
code, but is unlikely to stop a crash) then you could configure with

./configure CFLAGS=O0

(I think that's right - I confess that I always edit the src/Makefile
after the configure step to turn off the compiler optimizations).

Is what you are doing the same as that now documented at 
http://www.denemo.org/hacking-sources/#MacPorts
or do you have slightly different packages?

Richard



>  If I do a "make distclean" and
> try again without the "--enable-debug" then it segfaults. So this one
> change allows it to work.
> 
> I have no idea why this works, but it does.
> 
> Cheers
> 
> Eddy
> 
> 
> On Thu, Jun 25, 2020 at 10:41:54AM -0500, David Richmond wrote:
> > I do get the /proc/self/exe errors, but execution continues.
> > 
> > I compiled against gtk+quartz, by the way—no x11.
> > 
> > On Thu, Jun 25, 2020 at 10:21 AM Eddy <eddyjd97@gmail.com> wrote:
> > 
> > > I've tried on both 2.4.0 and git. It segfaults in the exact same
> > > place
> > > for both.
> > > 
> > > Do you not get the '/proc/self/exe: No such file or directory'
> > > errors
> > > when running it?
> > > 
> > > 
> > > 
> > > 
> > > On Thu, Jun 25, 2020 at 09:45:18AM -0500, David Richmond wrote:
> > > > I am very much not an expert on POSIX build systems, but I will
> > > > say that
> > > 
> > > I
> > > > downloaded the 2.4.0 source tarball (vs. cloning from git) and
> > > > there was
> > > 
> > > no
> > > > included autogen.sh. So as I recall I just ran mkdir bin && cd
> > > > bin then
> > > > ../configure, make, make install.
> > > > 
> > > > David
> > > > 
> > > > On Thu, Jun 25, 2020 at 8:32 AM Eddy <eddyjd97@gmail.com>
> > > > wrote:
> > > > 
> > > > > Hi David,
> > > > > 
> > > > > I'm using macports. The only change I made before compiling
> > > > > denemo was
> > > > > to autogen.sh, where I changed /usr/share/aclocal to
> > > > > /opt/local/share/aclocal.
> > > > > 
> > > > > Firing up denemo from lldb to see where the segfault is
> > > > > happening gives
> > > > > the following:
> > > > > 
> > > > >         * thread #1, queue = 'com.apple.main-thread', stop
> > > > > reason =
> > > > > EXC_BAD_ACCESS (code=1, address=0xfffffffffffffff8)
> > > > >                 frame #0: 0x000000010008eeac
> > > 
> > > denemo`init_lilypond_buffer
> > > > > at exportlilypond.c:3515 [opt]
> > > > >            3512       manager =
> > > > > gtk_source_language_manager_new ();
> > > > >            3513       gtk_source_language_manager_set_search_
> > > > > path
> > > > > (manager, (gchar**)dirs);
> > > > >            3514   }
> > > > >         -> 3515   const gchar * const *ids =
> > > > > gtk_source_language_manager_get_language_ids (manager);
> > > > >            3516
> > > > >            3517   GtkSourceLanguage *language =
> > > > > gtk_source_language_manager_get_language
> > > > > (manager,"lilypond");
> > > > >            3518   Denemo.textbuffer = (GtkTextBuffer *)
> > > > > gtk_source_buffer_new (tagtable);
> > > > >         Target 1: (denemo) stopped.
> > > > > 
> > > > > and the corresponding backtrace is:
> > > > > 
> > > > >         * thread #1, queue = 'com.apple.main-thread', stop
> > > > > reason =
> > > > > EXC_BAD_ACCESS (code=1, address=0xfffffffffffffff8)
> > > > >           * frame #0: 0x000000010008eeac
> > > > > denemo`init_lilypond_buffer at
> > > > > exportlilypond.c:3515 [opt]
> > > > >                 frame #1: 0x0000000100078b46
> > > > > denemo`inner_main(files=0x0000000000000000) at view.c:606
> > > > > [opt]
> > > > >                 frame #2: 0x00000001002d3437
> > > 
> > > libguile-2.2.1.dylib`c_body +
> > > > > 15
> > > > >                 frame #3: 0x00000001003411c2
> > > > > libguile-2.2.1.dylib`vm_regular_engine + 1477
> > > > >                 frame #4: 0x000000010033fe34
> > > > > libguile-2.2.1.dylib`scm_call_n + 380
> > > > >                 frame #5: 0x000000010033d75a
> > > 
> > > libguile-2.2.1.dylib`catch +
> > > > > 490
> > > > >                 frame #6: 0x00000001002d33c6
> > > > > libguile-2.2.1.dylib`scm_i_with_continuation_barrier + 137
> > > > >                 frame #7: 0x00000001002d341d
> > > > > libguile-2.2.1.dylib`scm_c_with_continuation_barrier + 60
> > > > >                 frame #8: 0x000000010033d2ad
> > > > > libguile-2.2.1.dylib`with_guile + 63
> > > > >                 frame #9: 0x00000001003e554c
> > > > > libgc.1.dylib`GC_call_with_stack_base + 26
> > > > >                 frame #10: 0x000000010033b81f
> > > > > libguile-2.2.1.dylib`scm_with_guile + 45
> > > > >                 frame #11: 0x0000000100069ad7
> > > > > denemo`main(argc=<unavailable>, argv=0x00007ffeefbff960) at
> > > > > main.c:452
> > > 
> > > [opt]
> > > > >                 frame #12: 0x00007fff6cfa5015
> > > > > libdyld.dylib`start + 1
> > > > > 
> > > > > Thanks for your help,
> > > > > 
> > > > > Eddy
> > > > > 
> > > > > 
> > > > > On Thu, Jun 25, 2020 at 07:07:54AM -0500, David Richmond
> > > > > wrote:
> > > > > > Hi Eddy,
> > > > > > 
> > > > > > I have Denemo 2.4.0 running successfully on Catalina
> > > > > > (10.15.4). What
> > > 
> > > did
> > > > > > you use to install the dependencies?
> > > > > > 
> > > > > > David
> > > > > > 
> 
>