[MPlayer-users] any success compiling with gcc-6 yet?

[Alexander Strasser]

Hi Arthur,

  thanks for testing build with gcc-6.

On 2015-12-20 08:46 +1030, Arthur Marsh wrote:
> Arthur Marsh wrote on 14/12/15 09:49:
> >Arthur Marsh wrote on 14/12/15 06:45:
> >>Hi, I've tried building mplayer with gcc-6 but it failed with a compiler
> >>error (which I can post if anyone's interested).
> >>
> >>Has anyone else been attempting to build mplayer with gcc-6?
> >>
> >>Arthur.
> >
> >libavcodec/pngdec.c:1086:12: internal compiler error: in contains_point,
                                ^^^^^^^^^^^^^^^^^^^^^^^
  
  I read ICE here...

> >at diagnostic-show-locus.c:340
> >  static int decode_frame_common(AVCodecContext *avctx, PNGDecContext *s,
> >             ^~~~~~~~~~~~~~~~~~~
> >
> >0x102bf6f contains_point
> >         ../../src/gcc/diagnostic-show-locus.c:340
> >0x102bf6f get_state_at_point
> >         ../../src/gcc/diagnostic-show-locus.c:684
> >0x102c518 print_source_line
> >         ../../src/gcc/diagnostic-show-locus.c:538
> >0x102c518 diagnostic_show_locus(diagnostic_context*, diagnostic_info
> >const*)
> >         ../../src/gcc/diagnostic-show-locus.c:803

  
  ...and from the remaining back trace one could assume
an error manifesting in the diagnostig reporting machinery
from gcc.

> >0x63e6d0 c_diagnostic_finalizer
> >         ../../src/gcc/c-family/c-opts.c:167
> >0x102a248 diagnostic_report_diagnostic(diagnostic_context*,
> >diagnostic_info*)
> >         ../../src/gcc/diagnostic.c:800
> >0x102b2f0 warning_at(unsigned int, int, char const*, ...)
> >         ../../src/gcc/diagnostic.c:1029
> >0xb3cfd1 warn_uninit
> >         ../../src/gcc/tree-ssa-uninit.c:157
> >0xb42446 warn_uninitialized_phi
> >         ../../src/gcc/tree-ssa-uninit.c:2366
> >0xb42446 execute
> >         ../../src/gcc/tree-ssa-uninit.c:2459
> >Please submit a full bug report,
> >with preprocessed source if appropriate.
> >Please include the complete backtrace with any bug report.
> >See <file:///usr/share/doc/gcc-6/README.Bugs> for instructions.
> >common.mak:60: recipe for target 'libavcodec/pngdec.o' failed

  libavcodec/pngdec.c is in FFmpeg, but as I said
before it very much looks like a bug in gcc.

> >building on Debian x86-64.
> 
> Another weird side-effect of running ./configure with --cc=gcc-6 is that
> dirac support gets removed.
> 
> I'll wait until Debian releases another snapshot unless anyone can suggest
> how to narrow down the bug into a reportable state.

  Maybe you should try reporting problems with gcc snapshots
from Debian to Debian first? In the hope they could handle
your report efficiently and maybe forward to gcc upstream
or e.g. send you to us if they think it is not a gcc problem.


  Alexander