[Home]

Summary:ASTERISK-00529: [request] [tracking] gsm codec won't build on PPC - error in Makefile
Reporter:iain (iain)Labels:
Date Opened:2003-11-15 11:29:22.000-0600Date Closed:2004-09-25 02:53:45
Priority:MinorRegression?No
Status:Closed/CompleteComponents:Core/General
Versions:Frequency of
Occurrence
Related
Issues:
Environment:Attachments:
Description:These lines in the Makefile:

 ifneq (${OSARCH},Darwin)
  OPTIMIZE+=-march=$(PROC)
 endif

result in arch being undefined on PPC - creating this error:

cc1: Invalid option "arch = "
Comments:By: Brian West (bkw918) 2003-11-19 19:11:22.000-0600

please attach a diff

By: iain (iain) 2003-11-22 06:24:22.000-0600

Problem is ... I don't know how to construct Makefiles.  So if I try and fix it I may break it for other platforms.  Any compile command that specifies -arch will fail on ppc so any line in the Makefile that specifies an architecture must default to not adding -arch

By: Tilghman Lesher (tilghman) 2003-12-12 00:32:25.000-0600

Prior to this line, include:

PROC=$(shell uname -m)

By: iain (iain) 2003-12-13 11:41:50.000-0600

Tried that :-(

It causes -march to be set to ppc but that is no help because the compilers shipped with ppc don't recognise the -march=ppc option.  The -march option must not be present at all for the codec to build.  This works:

ifneq (${OSARCH},Darwin)
ifneq ($(shell uname -m),ppc)
OPTIMIZE+=-march=$(PROC)
endif
endif

... is this a fix that will be OK on other platforms?

By: Brian West (bkw918) 2004-01-07 00:04:41.000-0600

try to attach a diff -u please.

By: iain (iain) 2004-01-09 16:05:07.000-0600

Apply against cvs as of 9/1/2004

--- asterisk/codecs/gsm/Makefile.orig   Fri Jan  9 21:41:14 2004
+++ asterisk/codecs/gsm/Makefile        Fri Jan  9 21:44:34 2004
@@ -39,7 +39,9 @@

ifneq (${OSARCH},Darwin)
ifneq (${PROC},x86_64)
+ifneq ($(shell uname -m),ppc)
OPTIMIZE+=-march=$(PROC)
+endif
endif
endif

That fixes the top level build.  However, it's still necessary to go to asterisk/codecs/gsm and fo a "make" otherwise I get a load of assembler error messages.

By: James Golovich (jamesgolovich) 2004-03-25 23:30:26.000-0600

This appears to have been fixed by the patch in bug ASTERISK-851.  This one should have been closed out long ago as duplicate