#414 Revert a temporary fix for mock >= 1.4.1 of year 2017
Merged 2 years ago by ngompa. Opened 3 years ago by sergiomb.
sergiomb/FedoraReview bootstrap-chroot  into  master

file modified
+1 -2
@@ -123,8 +123,7 @@ 

  .TP 4

  .B -o, --mock-options "options..."

  Mock options for the build. Defaults to --no-cleanup-after --no-clean,

- for use with mock >= 1.4.1 --no-bootstrap-chroot additionally, you might

- want this along with other options you provide.

+ you might want this along with other options you provide.

  .TP 4

  .B --no-report

  Do not generate the review template.

@@ -74,8 +74,6 @@ 

  def _mock_options_default():

      """ Returns the default mock options """

      mock_opts = "--no-cleanup-after --no-clean"

-     if version.parse(_check_mock_ver()) >= version.parse("1.4.1"):

-         mock_opts = "--no-bootstrap-chroot %s" % mock_opts

      return mock_opts

  

  

after first comit approved in the other PR , I'd like revert this temporary fix, to have a faster review

why this PR is ignored ?

this is fix a nasty bug , we shouldn't use --no-bootstrap-chroot , all builds take much longer

Sorry, the reason I haven't merged this is because I don't know what this is about. There is no explanation about what this is and what it's for.

Also, if we remove --no-bootstrap-chroot, we're going back to running two chroot setup cycles (one bootstrap and another regular), which would make this take longer.

rebased onto 48780d2

2 years ago

After review PR #413 , IIRC I notice this one , IIRC we shouldn't change default configuration , the original commit said temporary fix , which we don't need anymore I think , but I have to review it again

I remember that this patch increased revision time, but in today's tests it didn't, I think it's because today the mirror and the network are very fast..., but I don't see any utility in keeping the temporary fix, since the default settings of mock also work well

I tested with bug 1976038 (golang-github-sgreben-flagvar) [1] this time on first attempt with patch have bad results, but on second try I have much better results.

[1]

fedora-review -v -b 1976038
cp ~/.cache/fedora-review.log  fedora-review_with_patch.log

#removed the patch 
fedora-review -v -b 1976038
cp ~/.cache/fedora-review.log  fedora-review_without_patch.log
diff <(cut -b13- fedora-review_without_patch.log) <(cut -b13- fedora-review_with_patch.log) -up > test1.txt

#readd the patch
fedora-review -v -b 1976038
cp ~/.cache/fedora-review.log  fedora-review_with_patch_try2.log
diff <(cut -b13- fedora-review_without_patch.log) <(cut -b13- fedora-review_with_patch_try2.log) -up > teste2.txt

The chroots are cached. So after the first round, the overhead is close to zero. And it can save you a lot of headaches when new incompatibilities land in rawhide (mainly rpm).

@msuchy thank you for opinion , I think it is approved , please merge it , I had tested last 4 months ...

Pull-Request has been merged by ngompa

2 years ago