Build hangs on Solaris 10 SPARC with V1.3.0nb3 & gcc

[GoogleCodeExporter]

Hi,

I've never built this library before so please excuse my ignorance.

I'm using a pkgsrc build environment on Solaris 10 SPARC, gcc 4.5.2 and gmake. 
I've built quite a few packages with this setup so I don't think it's a problem 
with my environment. Here's what happens:

The compiler gets to this line and just hangs steadily reporting 85%+ cpu usage:

[CC] vp9/vp9_cx_iface.c.o
libtool: compile:  gcc -O2 
-I/home/isaleh/pkgsrc/multimedia/libvpx/work/.buildlink/include -O3 -fPIC -Wall 
-Wdeclaration-after-statement -Wdisabled-optimization -Wpointer-arith 
-Wtype-limits -Wcast-qual -Wvla -Wimplicit-function-declaration -Wuninitialized 
-Wunused-variable -Wno-unused-function -I. 
-I/home/isaleh/pkgsrc/multimedia/libvpx/work/libvpx-1.3.0 -c vp9/vp9_cx_iface.c 
 -fPIC -DPIC -o vp9/.libs/vp9_cx_iface.c.o

I've run that gcc build command outside outside of the make prcoess and it also 
just hangs. Truss (Solaris STRACE) doesn't seem to indicate that much is 
happening at this time that should be resource intensive.

I've let it run for several hours just for the heck of it and no change.

Any ideas why this is happening/how to fix it? 

Let me know which logs might give you the debugging info you need.

Thanks,
Ibraheem

Original issue reported on code.google.com by Shadowha...@gmail.com on 8 Apr 2015 at 4:22

[GoogleCodeExporter]

Bah... nevermind... Package src already has a workaround for this issue. I 
guess it didn't automatically apply it due to a version difference or something.

You can close this I guess.

Thanks!

Original comment by Shadowha...@gmail.com on 8 Apr 2015 at 8:51

[GoogleCodeExporter]

What needed to be changed to fix this? Do you know if it was pushed upstream 
and is available in 1.4.0?

Original comment by johannko...@google.com on 9 Apr 2015 at 3:29

[GoogleCodeExporter]

The pkgsrc hacks.mk file only says:
### GCC 4.5.3 gets stuck while compiline "vp9/vp9_cx_iface.c".
### See PR port-sparc64/48501 for more details.
...
PKG_HACKS+=             sparc-gcc-lock-up
BUILDLINK_TRANSFORM+=   rm:-O[0-9]*
...
I'm not entirely sure what's happening behind the scenes haha.
...
Here's the thread... It's a compiler bug which was fixed with GCC 4.8 it seems.
http://mail-index.netbsd.org/netbsd-bugs/2014/01/05/msg035549.html

Original comment by Shadowha...@gmail.com on 9 Apr 2015 at 10:50

[GoogleCodeExporter]

nvm... easy... it's just removing O flags :P

Original comment by Shadowha...@gmail.com on 9 Apr 2015 at 10:51

[GoogleCodeExporter]

Oh, it removes the '-On' optimization level? That's not something we will apply 
upstream :) sounds like it's also fixed with the newer compiler.

Original comment by johannko...@google.com on 9 Apr 2015 at 10:56

• Changed state: WontFix