[MacPorts] #31941: firefox-x11 7.0.1 fails to build

MacPorts noreply at macports.org
Thu Jan 19 02:24:07 PST 2012


#31941: firefox-x11 7.0.1  fails to build
-------------------------------+--------------------------------------------
 Reporter:  matuzalem@…        |       Owner:  jeremyhu@…           
     Type:  defect             |      Status:  new                  
 Priority:  Normal             |   Milestone:                       
Component:  ports              |     Version:  2.0.3                
 Keywords:                     |        Port:  firefox-x11          
-------------------------------+--------------------------------------------

Comment(by gary.verhaegen@…):

 I am now concentrating on the 32bit version; I would guess that the
 general principles are the same for the 64bit version, but there are fewer
 missing symbols in the 32bit version. I am not very comfortable with C
 programming, but I peeked around a bit and here's what I understand of
 this issue. At the start of the `state.c` file of `libtheora`, there are a
 bunch of includes based on architecture flags. The include for the `x86`
 architecture redefines the function `oc_state_accel_init_c` to be
 `oc_state_accel_init_x86`, which is declared in the included file
 (`x86int.h`) but not defined.

 Looking around, I did find a definition for `oc_state_accel_init_x86` in
 the `x86state.c` file (same folder as `x86int.h`: `mozilla-
 release/media/libtheora/lib/x86`). However, as there are no `.o` files in
 that directory, my guess would be that the build script somehow forgot to
 compile them and include them in the linking phase. Modifying the build
 script to correctly include these files is beyond me, but simply
 commenting out the redefinition of `oc_state_accel_init_c` to
 `oc_state_accel_init_x86` allowed me to finish the compilation. My guess
 is that my version of libtheora will be a bit slower, since it is using a
 `C` version instead of a hand-coded `x86` assembly version of that
 initialization function, but I hope everything will nonetheless keep
 working.

 So the problem seems to be with libtheora more than with firefox itself.
 As an aside, is there any reason for the firefox port to compile its own
 version of libtheora instead of relying on the existing `libtheora` port?
 (which compiles without any problem on my machine)

-- 
Ticket URL: <https://trac.macports.org/ticket/31941#comment:8>
MacPorts <http://www.macports.org/>
Ports system for Mac OS


More information about the macports-tickets mailing list