Description of problem: When running createrepo --compress-type=xz not all resulting compressed metadata is compressed with xz.
Version-Release number of selected component (if applicable):createrepo-0.9.9-23.el7
Steps to Reproduce:
1.createrepo --compress-type=xz .
3.files are present
some files are created with gzip compression
xz compression used for all files
__init__.py from createrepo has a 'FIXME' relating to y-m-p in the relevant code
While I was testing bug 874682 I did a note to the test that preserving gz format for primary metadata with --compress-type=xz was intentional but I cannot find the reason for that...
James or Valentina,
do you know why? Maybe we thought it could be risky for RHEL-6 since pyliblzma is not present in RHEL6.
Closest I've got for reasoning, but this doesn't truly clear your needinfo
(In reply to Karel Srot from comment #3)
> do you know why? Maybe we thought it could be risky for RHEL-6 since
> pyliblzma is not present in RHEL6.
It was simpler than that, at the time we passed the downloaded .xml files directly to yum-metadata-parser. And that used gzopen() if the file ended in .gz or open() otherwise.
Between el6 => el7 we changed yum so it decompresses the .xml files itself, and passes the result to yum-metadata-parser, avoiding this problem. So we could fix this in createrepo now.
Acking (fix hinted in comment 5).
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.
For information on the advisory, and where to find the updated
files, follow the link below.
If the solution does not work for you, open a new bug report.