Package: src:gcc-4.4 Version: 4.4.5-9 Severity: normal Tags: patch Hi, the attached .debdiff fixes PR44606 [1], a register-allocation bug that (seldomly) miscompiles floating point code in the unofficial powerpcspe port [2]. This is a backport of commit 168347 from GCC svn [3]. The patch modifies debian/rules.patch to only apply the patch on powerpcspe, so there should be no side-effects on other architectures. I verified the patch to compile and be effective on a powerpcspe sid installation running on a P2020RDB development board. I guess the same patch should go into gcc-4.3 as well, going to post a corresponding bug report in the next days. cheers, David [1] http://gcc.gnu.org/bugzilla/show_bug.cgi?id=44606 [2] http://wiki.debian.org/PowerPCSPEPort [3] http://gcc.gnu.org/viewcvs?view=revision&revision=168347 -- GnuPG public key: http://user.cs.tu-berlin.de/~dvdkhlng/dk.gpg Fingerprint: B17A DC95 D293 657B 4205 D016 7DEF 5323 C174 7D40
Attachment:
pr44606.debdiff
Description: Binary data
Attachment:
pgpZGp9EaqZwY.pgp
Description: PGP signature