Summary: | freetype (all versions) with gcc33 | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Dietrich Heise <dietrich> |
Component: | New packages | Assignee: | Gentoo Fonts Team <fonts> |
Status: | RESOLVED WONTFIX | ||
Severity: | enhancement | CC: | flash3001 |
Priority: | High | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
URL: | http://www.freetype.org/pipermail/devel/2003-June/009452.html | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: |
ftdump.c.patchgcc33
ftdump.c.patchgcc33 freetype-1.3.1-r3.ebuild |
Description
Dietrich Heise
2004-01-30 11:14:36 UTC
could you provide some further justification (eg, upstream) for this? I've never had any problems with freetype compiling at all. at most we should patch the makefiles to ensure the particular file are compiled in the right way OR let the upstream developers fix the code themselves. Anyway gcc is still smart enough to discover when the strict aliasing rules are to be avoided... Created attachment 24993 [details, diff]
ftdump.c.patchgcc33
patch freetype with with:
patch test/ftdump.c ftdump.c.patchgcc33
Created attachment 24995 [details, diff]
ftdump.c.patchgcc33
Created attachment 24996 [details]
freetype-1.3.1-r3.ebuild
i added -fno-strict-aliasing to 2.1.9, although for another reason. I see no reason to change this in all older versions as long as it doesn't break. Your report or the mail thread do not suggest that. What i do see is that disabling it is slower, that is not good. seems to be gone with gcc-3.4.x |