media-libs/svgalib-1.9.25 fails to build on linux kernel version 2.6.26. Errors are along the lines of undefuned variables, et al. Reproducible: Always Steps to Reproduce: 1. have kernel >= 2.6.26 2. emerge svgalib Actual Results: fails to build Expected Results: builds Fedora guys fixed this already... http://cvs.fedoraproject.org/viewvc/devel/svgalib/svgalib-1.9.25-kernel-2.6.26.patch?revision=1.2 I put this and the fixed ebuild (with the extra epatch) into my personal overlay - builds perfectly.
seems that this bug affects any program that uses lrmi - and only for x86. The code in the kernel that has changed is related to the VM86 code - http://www.ussg.iu.edu/hypermail/linux/kernel/0803.3/2188.html Other packages affected: so far, I've only found svgalib and read-edid, but I imagine there's more? Could a kernel patch be in order?
Thanks for the report. I don't think a patch backing out a purposeful change in kernel flags is a good idea. The programs using the old flags should be patched to handle the new flags, like the example patch you linked to. I'll go ahead and take this bug since there is no other svgalib maintainer currently; would you mind opening up another one for getting read-edid patched as well?
changed name to reflect that this is for svgalib Opening a new bug for read-edid.
i am encountering this bug under lrmi (itself) as well, can someone confirm, and look into this at the same time, because this may well be a case of stuff that is built against system lrmi rather than internal? attaching several install attempts of lrmi just encase it helps
Created attachment 163590 [details] lrmi 0.7, 0.8, 0.10, 0.10-r1 and svgalib 1.9.25
(In reply to comment #5) > Created an attachment (id=163590) [edit] > lrmi 0.7, 0.8, 0.10, 0.10-r1 and svgalib 1.9.25 > (In reply to comment #4) > i am encountering this bug under lrmi (itself) as well, can someone confirm, > and look into this at the same time, because this may well be a case of stuff > that is built against system lrmi rather than internal? attaching several > install attempts of lrmi just encase it helps > It's lrmi that has this problem, true. However, svgalib and read-edid have their own modified lrmi implementations in their source trees. See bug 235495 for lrmi fix
thanks for lmri info, i tried to find the lmri filed bug, but there were no results so thanks for the link
Fixed in CVS using patch from bug 235495