Page MenuHomeSoftware Heritage
Differential D3306

loader: Read snapshot out of the last origin visit status
ClosedPublic
Actions

Authored by ardumont on Jun 17 2020, 4:26 PM.

Details

Reviewers
douardda
Group Reviewers
Reviewers
Maniphest Tasks
T2310: Make origin visits immutable
Commits
rDLDHGd210f73957a5: loader: Read snapshot out of the last origin visit status
Summary

Preparatory work to drop unused fields from OriginVisit

Related to T2310

Test Plan

tox

Diff Detail

Repository
rDLDHG Mercurial loader
Branch
master
Lint
No Linters Available
Unit
No Unit Test Coverage
Build Status
Buildable 13021
Build 19847: Phabricator diff pipeline on jenkinsJenkins console · Jenkins
Build 19846: arc lint + arc unit

Event Timeline

ardumont created this revision.Jun 17 2020, 4:26 PM
Herald added a reviewer: Reviewers. · View Herald TranscriptJun 17 2020, 4:26 PM
swh-public-ci added a comment.Jun 17 2020, 4:28 PM

Build is green

Patch application report for D3306 (id=11710)

Rebasing onto f1866671a4...

Current branch diff-target is up to date.
Changes applied before test
commit 63957ffc0cba296c37bc5f835a0d011ffa2e65e2
Author: Antoine R. Dumont (@ardumont) <ardumont@softwareheritage.org>
Date:   Wed Jun 17 16:22:12 2020 +0200

    loader: Read snapshot out of the last origin visit status
    
    Related to T2310

See https://jenkins.softwareheritage.org/job/DLDHG/job/tests-on-diff/9/ for more details.

Harbormaster completed remote builds in B12919: Diff 11710.Jun 17 2020, 4:28 PM
ardumont updated this revision to Diff 11711.Jun 17 2020, 4:33 PM

Forgot to update test accordingly

swh-public-ci added a comment.Jun 17 2020, 4:35 PM

Build is green

Patch application report for D3306 (id=11711)

Rebasing onto f1866671a4...

Current branch diff-target is up to date.
Changes applied before test
commit 151831ca932d85a2a75b93951f4245e3027f5c14
Author: Antoine R. Dumont (@ardumont) <ardumont@softwareheritage.org>
Date:   Wed Jun 17 16:22:12 2020 +0200

    loader: Read snapshot out of the last origin visit status
    
    Related to T2310

See https://jenkins.softwareheritage.org/job/DLDHG/job/tests-on-diff/10/ for more details.

Harbormaster completed remote builds in B12920: Diff 11711.Jun 17 2020, 4:35 PM
ardumont mentioned this in T2310: Make origin visits immutable.Jun 17 2020, 5:31 PM
ardumont planned changes to this revision.Jun 18 2020, 7:48 AM
ardumont mentioned this in D3313: algos.origin: Open origin-get-latest-visit-status function.Jun 18 2020, 1:45 PM
ardumont mentioned this in D3305: package.loader: Retrieve latest snapshot out of the latest visit status.Jun 19 2020, 3:40 PM
ardumont updated this revision to Diff 11806.Jun 22 2020, 4:20 PM

Simplify according to latest storage algos.snapshot.snapshot-get-latest
endpoint

swh-public-ci added a comment.Jun 22 2020, 4:21 PM

Build is green

Patch application report for D3306 (id=11806)

Rebasing onto cba09ab734...

Current branch diff-target is up to date.
Changes applied before test
commit 7837917c8ab71aefa1ad3da816defb900668f319
Author: Antoine R. Dumont (@ardumont) <ardumont@softwareheritage.org>
Date:   Mon Jun 22 16:02:57 2020 +0200

    loader: Read snapshot out of the last origin visit status
    
    Related to T2310

See https://jenkins.softwareheritage.org/job/DLDHG/job/tests-on-diff/11/ for more details.

Harbormaster completed remote builds in B13021: Diff 11806.Jun 22 2020, 4:21 PM
ardumont added inline comments.Jun 22 2020, 4:25 PM
swh/loader/mercurial/loader.py
173

I could require branches_count=1 as well as we don't do much with that snapshot.
Currently that call could end up being heavy if the snapshot is huge (and in the end we only use it to have its id...

ardumont updated this revision to Diff 11808.Jun 22 2020, 4:25 PM

Make it apparent we only use the snapshot for its id
(Also reduce the snapshot object model to the minimum)

swh-public-ci added a comment.Jun 22 2020, 4:27 PM

Build is green

Patch application report for D3306 (id=11808)

Rebasing onto cba09ab734...

Current branch diff-target is up to date.
Changes applied before test
commit 33fe79eb34c074a760fae5a67f2130e1d47ea7c9
Author: Antoine R. Dumont (@ardumont) <ardumont@softwareheritage.org>
Date:   Mon Jun 22 16:02:57 2020 +0200

    loader: Retrieve last snapshot with snapshot_get_latest function
    
    Related to T2310

See https://jenkins.softwareheritage.org/job/DLDHG/job/tests-on-diff/12/ for more details.

Harbormaster completed remote builds in B13023: Diff 11808.Jun 22 2020, 4:27 PM
ardumont added inline comments.Jun 22 2020, 4:27 PM
swh/loader/mercurial/loader.py
173

well, i did ;)
shout out if you don't agree.

douardda accepted this revision.Jun 23 2020, 9:42 AM
douardda added a subscriber: douardda.

Not sure about implications of the branches_count=1 argument, but it may be related to my other comment. Otherwise, LGTM.

swh/loader/mercurial/loader.py
602

Is there a corner case situation here where snapshots[-1] is a partial snapshot (e.g. because of a failure in the middle of the previous loading) but snapshots[-2] is ok, and current snapshot is actually the same this later?

In other words, should we check for existence of a previous snapshot equal to the current one, and not only the last one?

This revision is now accepted and ready to land.Jun 23 2020, 9:42 AM
ardumont added inline comments.Jun 23 2020, 9:54 AM
swh/loader/mercurial/loader.py
602

I guess that could happen yes.
Note that i'm just porting the code here ;)

We could adapt to iterate over all the snapshots we know for the origin.
(And i guess that could also be a common behavior for all loaders.)

I don't really know though:

  • if that situation happens enough
  • how complicate the code will get
  • if the worse situation case (no snapshot found over a lot of visits) how the performance will get degraded

All in all, i found the current code simple enough and to the point for most cases.
But as usual, i may be wrong ;)

ardumont added a comment.Jun 23 2020, 9:55 AM

Not sure about implications of the branches_count=1 argument, but it may be related to my other comment. Otherwise, LGTM.

As we only want the snapshot id here, there is no point in creating the full snapshot here (which in some edge cases could be large).
That's what the branches_count=1 is all about.

ardumont updated this revision to Diff 11827.Jun 23 2020, 11:36 AM

loader: Read snapshot out of the last origin visit status

swh-public-ci added a comment.Jun 23 2020, 11:37 AM

Build is green

Patch application report for D3306 (id=11827)

Rebasing onto cba09ab734...

Current branch diff-target is up to date.
Changes applied before test
commit d210f73957a52888ba6e17e118fa4e4485e3536b
Author: Antoine R. Dumont (@ardumont) <ardumont@softwareheritage.org>
Date:   Mon Jun 22 16:02:57 2020 +0200

    loader: Read snapshot out of the last origin visit status
    
    Related to T2310

See https://jenkins.softwareheritage.org/job/DLDHG/job/tests-on-diff/13/ for more details.

Harbormaster completed remote builds in B13041: Diff 11827.Jun 23 2020, 11:37 AM
Closed by commit rDLDHGd210f73957a5: loader: Read snapshot out of the last origin visit status (authored by ardumont). · Explain WhyJun 23 2020, 2:28 PM
This revision was automatically updated to reflect the committed changes.
ardumont added a commit: rDLDHGd210f73957a5: loader: Read snapshot out of the last origin visit status.

Revision Contents
Changeset List

PathSizeCoverage (All)Coverage (Touched)
swh/
loader/
mercurial/
8 lines90%
Loading...

Diff 11806

View Options

swh/loader/mercurial/loader.py

Loading...
About · Developer information · Legal