On Fri, 2018-09-28 at 11:01 -0700, Zac Medico wrote: > On 09/28/2018 12:55 AM, Michał Górny wrote: > > Whenever the install directory contains files that would collide upon > > (re)compressing, report them explicitly and fail. This indicates > > a serious problem in ebuild and since we don't know which of the files > > is correct, we should not attempt to choose between them. > > > > To reduce performance impact, the check is only done whenever compressed > > files are found. This is sufficient since for issue to occur there must > > be at least one compressed variant. > > > > Bug: https://bugs.gentoo.org/667072 > > Signed-off-by: Michał Górny <mgo...@gentoo.org> > > --- > > bin/ecompress | 24 ++++++++++++++++++++++++ > > 1 file changed, 24 insertions(+) > > > > diff --git a/bin/ecompress b/bin/ecompress > > index 36bdb585b..bc1f5e08a 100755 > > --- a/bin/ecompress > > +++ b/bin/ecompress > > @@ -49,6 +49,30 @@ while [[ $# -gt 0 ]] ; do > > find_args+=( -size > > "+${PORTAGE_DOCOMPRESS_SIZE_LIMIT}c" ) > > > > while IFS= read -d '' -r path; do > > + # detect the horrible posibility of the ebuild > > installing > > + # colliding compressed and/or uncompressed > > variants > > + # and fail hard (bug #667072) > > + # > > + # note: to save time, we need to do this only > > if there's > > + # at least one compressed file > > + case ${path} in > > + *.Z|*.gz|*.bz2|*.lzma|*.xz) > > + vpath=${path%.*} > > + for comp in '' .Z .gz .bz2 > > .lzma .xz; do > > + if [[ ${vpath}${comp} > > != ${path} && \ > > + -e > > ${vpath}${comp} ]]; then > > + eerror > > "Colliding files found for ecompress:" > > + eerror > > + eerror " > > ${path#${D%/}}" > > + eerror " > > ${vpath#${D%/}}${comp}" > > + eerror > > + eerror "Please > > remove the incorrect of those files." > > + die "Aborting > > due to colliding compressed files." > > + fi > > + done > > + ;; > > + esac > > + > > >> "${path}.ecompress" || die > > done < <(find "${find_args[@]}" -print0 || die) > > fi > > > > This breaks compatibility, so let's use eqawarn and simply leave the > compressed files in place for now, then make it die in the next EAPI.
Compatibility with what? With one ebuild where maintainer failed to handle recent eqawarn? Given that he couldn't have expected any sane behavior previously? -- Best regards, Michał Górny
signature.asc
Description: This is a digitally signed message part
