1 |
On Tue, 2018-09-25 at 12:58 -0700, Zac Medico wrote: |
2 |
> On 09/25/2018 12:55 PM, Zac Medico wrote: |
3 |
> > On 09/25/2018 12:18 PM, Michał Górny wrote: |
4 |
> > > Fix crash due to race condition in handling the same file being present |
5 |
> > > both in compressed and uncompressed variants. If that is the case, |
6 |
> > > just queue the uncompressed variant for compression, and ignore |
7 |
> > > the other compressed variants. |
8 |
> > > |
9 |
> > > Bug: https://bugs.gentoo.org/667072 |
10 |
> > > Signed-off-by: Michał Górny <mgorny@g.o> |
11 |
> > > --- |
12 |
> > > bin/ecompress | 7 +++++++ |
13 |
> > > 1 file changed, 7 insertions(+) |
14 |
> > > |
15 |
> > > diff --git a/bin/ecompress b/bin/ecompress |
16 |
> > > index 36bdb585b..d5ff3796c 100755 |
17 |
> > > --- a/bin/ecompress |
18 |
> > > +++ b/bin/ecompress |
19 |
> > > @@ -49,6 +49,13 @@ while [[ $# -gt 0 ]] ; do |
20 |
> > > find_args+=( -size "+${PORTAGE_DOCOMPRESS_SIZE_LIMIT}c" ) |
21 |
> > > |
22 |
> > > while IFS= read -d '' -r path; do |
23 |
> > > + # if both compressed and uncompressed variant exists, |
24 |
> > > + # skip the compressed variants (bug #667072) |
25 |
> > > + case ${path} in |
26 |
> > > + *.Z|*.gz|*.bz2|*.lzma|*.xz) |
27 |
> > > + [[ -s ${path%.*} ]] && continue |
28 |
> > > + ;; |
29 |
> > > + esac |
30 |
> > |
31 |
> > In theory, we'd still have a problem if the file existed with muliple |
32 |
> > compressions, right? Maybe a good solution is to strip the compression |
33 |
> > extension here, and then have ecompress-file check for duplicates? |
34 |
> |
35 |
> Alternatively, we could uncompress the pre-compressed files right here. |
36 |
|
37 |
That would cause them to be decompressed even if the path is eventually |
38 |
exclude via 'docompress -x'. |
39 |
|
40 |
-- |
41 |
Best regards, |
42 |
Michał Górny |