When using INSTALL_MASK="*.a" for com_err the file "libcom_err.a" is not installed and as a later result the sys-fs/e2fsprogs wrongly fails. This was done to fix bug #125146 as you can see. [snip] pkg_setup() { # sanity check for #125146 if [[ -L ${ROOT}/usr/$(get_libdir)/libcom_err.a ]] || \ [[ ! -e ${ROOT}/usr/$(get_libdir)/libcom_err.a ]] then rm -f "${ROOT}"/usr/$(get_libdir)/libcom_err.a eerror "Your libcom_err.a is broken, please re-emerge com_err:" eerror " # emerge com_err" die "Mr. T pities the fool with a broken libcom_err.a" fi } [/snip] Here's the actual error. >>> Emerging (27 of 51) sys-fs/e2fsprogs-1.39 to /newroot/ >>> Extracting info * Your libcom_err.a is broken, please re-emerge com_err: * # emerge com_err !!! ERROR: sys-fs/e2fsprogs-1.39 failed. Call stack: ebuild.sh, line 1630: Called dyn_setup ebuild.sh, line 702: Called qa_call 'pkg_setup' ebuild.sh, line 38: Called pkg_setup e2fsprogs-1.39.ebuild, line 31: Called die !!! Mr. T pities the fool with a broken libcom_err.a !!! If you need support, post the topmost build error, and the call stack if relevant. !!! A complete build log is located at '/var/tmp/portage/sys-fs/e2fsprogs-1.39/temp/build.log'. !!! Setup failed: 1 Stripping libcom_err.a can be valid in some cases (embedded systems) and shouldn't cause e2fsprogs to die. Reproducible: Always Steps to Reproduce: 1. INSTALL_MASK="*.a" emerge com_err 2. emerge e2fsprogs Actual Results: Package unexpectedly dies Expected Results: Correctly handles behaviour if libcom_err.a is not present
that check is no longer needed since we know com_err-1.39 is not broken ... however, i'm not entirely sure how you're abusing ROOT here, so dont expect support if: - you're installing into $ROOT - you're trying to compile new applications against $ROOT in other words, you should not mix your development root and your target rootfs
