Created attachment 326450 [details, diff] ebuild.patch This patch fixes bug 243866
Created attachment 326452 [details] new ebuild
Comment on attachment 326450 [details, diff] ebuild.patch >- cd "${WORKDIR}"/binutils-${BASE_BINUTILS} >+ cd "${WORKDIR}"/binutils-${BASE_BINUTILS} || die i don't see much point to adding `die` to any of these cd commands. if you failed to change to the correct dir, follow up commands would all fail. >+ cd "${WORKDIR}"/binutils-${BASE_BINUTILS} || die >+ for i in .brik md5.sum INSTALL maintainer-scripts binutils cpu etc gas/doc gas/po gprof include/nlm include/regs ld texinfo >+ do >+ rm -fr "${i}" >+ done ignoring the problems with the code/style, there's no need for a for loop. pass all the paths to a single `rm -rf`. >+ mkdir -p "${WORKDIR}"/build/binutils || die >+ mkdir "${WORKDIR}"/build/gcc || die do it in a single call, and drop the die since it's pretty pointless with -p: mkdir -p "${WORKDIR}"/build/{binutils,gcc} >+ tc-export CC AR RANLIB i don't think this is the way to fix it. don't run configure directly ... use econf and set CTARGET=m68k-coff when doing so. >+ for i in a68k ld-tigcc tigcc/src tprbuilder/src patcher/src >+ do you have to declare new local vars as local, and the "do" should be on the previous line >+ cd "${S}"/sources/${i} || die >+ emake -e i don't think we want -e to emake. you also don't need the `cd` if you do: emake -C "${S}"/sources/${i}
+*tigcc-0.96_beta8-r1 (25 Dec 2014) + + 25 Dec 2014; Sergey Popov <pinkbyte@gentoo.org> -tigcc-0.96_beta8.ebuild, + +tigcc-0.96_beta8-r1.ebuild: + Revision bump: respect AR, CC and RANLIB, wrt bug #243866. Bump EAPI and fix + some QA issues, wrt bug #438218. Acked by Mikle Kolyada. Drop old revision