[SCM] lame packaging branch, karmic, updated. debian/3.98.2+debian-0ubuntu2
siretart at users.alioth.debian.org
siretart at users.alioth.debian.org
Sat Oct 24 16:11:33 UTC 2009
The following commit has been merged in the karmic branch:
commit 10b4a3cbbfe25e4931455ef995200ce32ae6df31
Author: Reinhard Tartler <siretart at tauware.de>
Date: Thu Oct 22 22:14:50 2009 +0200
buggy optimization for single-precision floats when -ffinite-math-only
From the marillat package:
lame (3.98.2-0.4) unstable; urgency=low
* Added a patch from Martin Guy to fix a silent output file on armel arch.
-- Christian Marillat <marillat at debian.org> Fri, 20 Mar 2009 09:30:45 +0100
diff --git a/debian/patches/05-armel-no-finite-math-only.diff b/debian/patches/05-armel-no-finite-math-only.diff
new file mode 100644
index 0000000..b99402c
--- /dev/null
+++ b/debian/patches/05-armel-no-finite-math-only.diff
@@ -0,0 +1,44 @@
+A bug present in gcc-4.[123] up to 4.3.3 makes a buggy optimization for
+single-precision floats when -ffinite-math-only (enabled by -ffast-math)
+
+Symptom: lame -V0 produces silent output files.
+Refs:
+https://sourceforge.net/tracker/index.php?func=detail&aid=2681646&group_id=290&atid=100290
+https://trac.xiph.org/ticket/1526
+http://gcc.gnu.org/bugzilla/show_bug.cgi?id=39501
+
+ Martin Guy <martinwguy at yahoo.it> 19 March 2009
+
+--- lame-3.98.2/configure.old 2008-06-29 22:40:05.000000000 +0100
++++ lame-3.98.2/configure 2009-03-19 19:00:42.000000000 +0000
+@@ -30189,6 +30189,15 @@
+ CFLAGS="-Wall ${CFLAGS}"
+ fi
+
++ # Work round buggy softfloat optimization in ARM EABI compilers
++ # -gnueabi in only gcc-4.1 onwards
++ FAST_MATH="-ffast-math"
++ case "${host}" in
++ arm*-gnueabi)
++ FAST_MATH="-ffast-math -fno-finite-math-only"
++ ;;
++ esac
++
+ # GCC version specific generic options
+ case "${GCC_version}" in
+ 2.96*)
+@@ -30217,12 +30226,12 @@
+ ;;
+ esac
+
+- OPTIMIZATION="-O3 ${OMIT_FRAME_POINTER} -ffast-math"
++ OPTIMIZATION="-O3 ${OMIT_FRAME_POINTER} ${FAST_MATH}"
+ OPTIMIZATION_FULL="-fmove-all-movables -freduce-all-givs \
+ -fbranch-count-reg -fforce-addr -fforce-mem"
+ ;;
+ 4.*)
+- OPTIMIZATION="-O3 -fomit-frame-pointer -ffast-math"
++ OPTIMIZATION="-O3 -fomit-frame-pointer ${FAST_MATH}"
+ OPTIMIZATION_FULL="-fbranch-count-reg -fforce-addr"
+ ;;
+ *)
diff --git a/debian/patches/05-armel-no-finite-math-only.patch b/debian/patches/05-armel-no-finite-math-only.patch
new file mode 100644
index 0000000..b99402c
--- /dev/null
+++ b/debian/patches/05-armel-no-finite-math-only.patch
@@ -0,0 +1,44 @@
+A bug present in gcc-4.[123] up to 4.3.3 makes a buggy optimization for
+single-precision floats when -ffinite-math-only (enabled by -ffast-math)
+
+Symptom: lame -V0 produces silent output files.
+Refs:
+https://sourceforge.net/tracker/index.php?func=detail&aid=2681646&group_id=290&atid=100290
+https://trac.xiph.org/ticket/1526
+http://gcc.gnu.org/bugzilla/show_bug.cgi?id=39501
+
+ Martin Guy <martinwguy at yahoo.it> 19 March 2009
+
+--- lame-3.98.2/configure.old 2008-06-29 22:40:05.000000000 +0100
++++ lame-3.98.2/configure 2009-03-19 19:00:42.000000000 +0000
+@@ -30189,6 +30189,15 @@
+ CFLAGS="-Wall ${CFLAGS}"
+ fi
+
++ # Work round buggy softfloat optimization in ARM EABI compilers
++ # -gnueabi in only gcc-4.1 onwards
++ FAST_MATH="-ffast-math"
++ case "${host}" in
++ arm*-gnueabi)
++ FAST_MATH="-ffast-math -fno-finite-math-only"
++ ;;
++ esac
++
+ # GCC version specific generic options
+ case "${GCC_version}" in
+ 2.96*)
+@@ -30217,12 +30226,12 @@
+ ;;
+ esac
+
+- OPTIMIZATION="-O3 ${OMIT_FRAME_POINTER} -ffast-math"
++ OPTIMIZATION="-O3 ${OMIT_FRAME_POINTER} ${FAST_MATH}"
+ OPTIMIZATION_FULL="-fmove-all-movables -freduce-all-givs \
+ -fbranch-count-reg -fforce-addr -fforce-mem"
+ ;;
+ 4.*)
+- OPTIMIZATION="-O3 -fomit-frame-pointer -ffast-math"
++ OPTIMIZATION="-O3 -fomit-frame-pointer ${FAST_MATH}"
+ OPTIMIZATION_FULL="-fbranch-count-reg -fforce-addr"
+ ;;
+ *)
diff --git a/debian/patches/series b/debian/patches/series
index 28cc3e6..437c44f 100644
--- a/debian/patches/series
+++ b/debian/patches/series
@@ -2,3 +2,4 @@
02-dynamic-frontends.patch
03-buffer-too-small.patch
04-exec-stac.patch
+05-armel-no-finite-math-only.patch
\ No newline at end of file
--
lame packaging
More information about the pkg-multimedia-commits
mailing list