Summary: | dev-scheme/drscheme miscompiles with -mfpmath=sse | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Michal Suchanek <hramrach> |
Component: | Current packages | Assignee: | Scheme Project <scheme> |
Status: | RESOLVED WONTFIX | ||
Severity: | major | CC: | bugs+gentoo, isaac.chanin, Michael.Gaber |
Priority: | High | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: | updatet ebuild with flag-o-matic removing of -mafpmath=sse |
Description
Michal Suchanek
2007-05-12 20:51:14 UTC
Have you tried recompiling without -mfpmath=sse in your cflags? Removing that flag fixed a rather similar segfault (and similar backtrace) for me. Though, as I'm using a prescott chip, your luck may vary. can confirm this problem and also the "solution" on a core2-duo (-march=prescott) and drscheme-371 Created attachment 133701 [details]
updatet ebuild with flag-o-matic removing of -mafpmath=sse
Michael, you're on i686? i don't get what you want to know with your question but atm my CFLAGS look like CFLAGS="-O2 -march=prescott -fomit-frame-pointer -pipe -msse -msse2 -msse3 -mmmx -mfpmath=sse" (In reply to comment #5) > i don't get what you want to know with your question I want to know your CHOST, the arch you are running. I want to know that it is not amd64. it is not amd64 Can anyone reproduce with 371? (In reply to comment #8) > Can anyone reproduce with 371? > Still segfaults with -mfpmath=sse and works fine without it, for me. |