| Submitter | Laurent Charignon |
|---|---|
| Date | May 12, 2016, 1:15 p.m. |
| Message ID | <05aaad2606ba0f2c6a0b.1463058915@dev5073.prn1.facebook.com> |
| Download | mbox | patch |
| Permalink | /patch/15076/ |
| State | Accepted |
| Headers | show |
Comments
> This bug uncovered another but (sp) in histedit
Correct, and this version of the fix does not require any change to the histedit test as it does not exercise the potentially buggy code path. From: <timeless.bmo1@gmail.com<mailto:timeless.bmo1@gmail.com>> on behalf of timeless <timeless@gmail.com<mailto:timeless@gmail.com>> Reply-To: "timeless@gmail.com<mailto:timeless@gmail.com>" <timeless@gmail.com<mailto:timeless@gmail.com>> Date: Thursday, May 12, 2016 at 3:15 PM To: Laurent Charignon <lcharignon@fb.com<mailto:lcharignon@fb.com>> Cc: mercurial-devel <mercurial-devel@mercurial-scm.org<mailto:mercurial-devel@mercurial-scm.org>> Subject: Re: [PATCH V2] strip: invalidate phase cache after stripping changeset (issue5235) > This bug uncovered another but (sp) in histedit
This patch looks pretty straight forward and sane to me. On 5/12/16 9:15 AM, Laurent Charignon wrote: > # HG changeset patch > # User Laurent Charignon <lcharignon@fb.com> > # Date 1463058839 25200 > # Thu May 12 06:13:59 2016 -0700 > # Node ID 05aaad2606ba0f2c6a0bf088d528dacf1436c419 > # Parent df838803c1d487e4601f96c6cfd85e6ad4f6291f > strip: invalidate phase cache after stripping changeset (issue5235) > > When we remove a changeset from the changelog, the phase cache must be > invalidated, otherwise it could refer to changesets that are no longer in the > repo. > > To reproduce the failure, I created an extension querying the phase cache after > the strip transaction is over. > > To do that, I stripped two commits with a bookmark on one of them to force > another transaction (we open a transaction for moving bookmarks) > after the strip transaction. > > Without the fix in this patch, the test leads to a stacktrace showing the issue: > repair.strip(ui, repo, revs, backup) > File "/Users/lcharignon/facebook-hg-rpms/hg-crew/mercurial/repair.py", line 205, in strip > tr.close() > File "/Users/lcharignon/facebook-hg-rpms/hg-crew/mercurial/transaction.py", line 44, in _active > return func(self, *args, **kwds) > File "/Users/lcharignon/facebook-hg-rpms/hg-crew/mercurial/transaction.py", line 490, in close > self._postclosecallback[cat](self) > File "$TESTTMP/crashstrip2.py", line 4, in test > [repo.changelog.node(r) for r in repo.revs("not public()")] > File "/Users/lcharignon/facebook-hg-rpms/hg-crew/mercurial/changelog.py", line 337, in node > return super(changelog, self).node(rev) > File "/Users/lcharignon/facebook-hg-rpms/hg-crew/mercurial/revlog.py", line 377, in node > return self.index[rev][7] > IndexError: revlog index out of range > > The situation was encountered in inhibit (evolve's repo) where we would crash > following the volatile set invalidation submitted by Augie in > e6f490e328635312ee214a12bc7fd3c7d46bf9ce. Before his patch the issue was masked > as we were not accessing the phasecache after stripping a revision. > > This bug uncovered another but in histedit (see explanation in issue5235). > I changed the histedit test accordingly to avoid fixing two things at once. > > diff --git a/mercurial/repair.py b/mercurial/repair.py > --- a/mercurial/repair.py > +++ b/mercurial/repair.py > @@ -194,6 +194,7 @@ > if not repo.ui.verbose: > repo.ui.popbuffer() > f.close() > + repo._phasecache.invalidate() > > for m in updatebm: > bm[m] = repo[newbmtarget].node() > diff --git a/tests/test-strip.t b/tests/test-strip.t > --- a/tests/test-strip.t > +++ b/tests/test-strip.t > @@ -838,6 +838,41 @@ > date: Thu Jan 01 00:00:00 1970 +0000 > summary: mergeCD > > +Check that the phase cache is properly invalidated after a strip with bookmark. > + > + $ cat > ../stripstalephasecache.py << EOF > + > from mercurial import extensions, localrepo > + > def transactioncallback(orig, repo, desc, *args, **kwargs): > + > def test(transaction): > + > # observe cache inconsistency > + > try: > + > [repo.changelog.node(r) for r in repo.revs("not public()")] > + > except IndexError: > + > repo.ui.status("Index error!\n") > + > transaction = orig(repo, desc, *args, **kwargs) > + > # warm up the phase cache > + > list(repo.revs("not public()")) > + > if desc != 'strip': > + > transaction.addpostclose("phase invalidation test", test) > + > return transaction > + > def extsetup(ui): > + > extensions.wrapfunction(localrepo.localrepository, "transaction", > + > transactioncallback) > + > EOF > + $ hg up -C 2 > + 1 files updated, 0 files merged, 0 files removed, 0 files unresolved > + $ echo k > k > + $ hg add k > + $ hg commit -m commitK > + $ echo l > l > + $ hg add l > + $ hg commit -m commitL > + $ hg book -r tip blah > + $ hg strip ".^" --config extensions.crash=$TESTTMP/stripstalephasecache.py > + 0 files updated, 0 files merged, 2 files removed, 0 files unresolved > + saved backup bundle to $TESTTMP/issue4736/.hg/strip-backup/8f0b4384875c-4fa10deb-backup.hg (glob) > + $ hg up -C 1 > + 0 files updated, 0 files merged, 1 files removed, 0 files unresolved > > Error during post-close callback of the strip transaction > (They should be gracefully handled and reported) > _______________________________________________ > Mercurial-devel mailing list > Mercurial-devel@mercurial-scm.org > https://urldefense.proofpoint.com/v2/url?u=https-3A__www.mercurial-2Dscm.org_mailman_listinfo_mercurial-2Ddevel&d=CwIGaQ&c=5VD0RTtNlTh3ycd41b3MUw&r=nuarHzhP1wi1T9iURRCj1A&m=lIZfwj_e7hmKsNbSd9ENLDYtrbWZUnuHNDwQlvjQWr8&s=p3QJeWzAtFbhXlCbuVLgJsIjVJpMJPhbz6iEfHypPlo&e=
On Thu, 19 May 2016 15:00:48 -0400, Durham Goode wrote: > This patch looks pretty straight forward and sane to me. LGTM, too. Should I move it to stable? > On 5/12/16 9:15 AM, Laurent Charignon wrote: > > # HG changeset patch > > # User Laurent Charignon <lcharignon@fb.com> > > # Date 1463058839 25200 > > # Thu May 12 06:13:59 2016 -0700 > > # Node ID 05aaad2606ba0f2c6a0bf088d528dacf1436c419 > > # Parent df838803c1d487e4601f96c6cfd85e6ad4f6291f > > strip: invalidate phase cache after stripping changeset (issue5235)
On 5/22/16 2:35 PM, Yuya Nishihara wrote: > On Thu, 19 May 2016 15:00:48 -0400, Durham Goode wrote: >> This patch looks pretty straight forward and sane to me. > LGTM, too. Should I move it to stable? I'm not sure how the new review system works, but yea. I vote push it. >> On 5/12/16 9:15 AM, Laurent Charignon wrote: >>> # HG changeset patch >>> # User Laurent Charignon <lcharignon@fb.com> >>> # Date 1463058839 25200 >>> # Thu May 12 06:13:59 2016 -0700 >>> # Node ID 05aaad2606ba0f2c6a0bf088d528dacf1436c419 >>> # Parent df838803c1d487e4601f96c6cfd85e6ad4f6291f >>> strip: invalidate phase cache after stripping changeset (issue5235)
On Mon, 23 May 2016 18:14:37 +0100, Durham Goode wrote: > On 5/22/16 2:35 PM, Yuya Nishihara wrote: > > On Thu, 19 May 2016 15:00:48 -0400, Durham Goode wrote: > >> This patch looks pretty straight forward and sane to me. > > LGTM, too. Should I move it to stable? > I'm not sure how the new review system works, but yea. I vote push it. Pushed to the stable branch of the committed repo, thanks.
Patch
diff --git a/mercurial/repair.py b/mercurial/repair.py --- a/mercurial/repair.py +++ b/mercurial/repair.py @@ -194,6 +194,7 @@ if not repo.ui.verbose: repo.ui.popbuffer() f.close() + repo._phasecache.invalidate() for m in updatebm: bm[m] = repo[newbmtarget].node() diff --git a/tests/test-strip.t b/tests/test-strip.t --- a/tests/test-strip.t +++ b/tests/test-strip.t @@ -838,6 +838,41 @@ date: Thu Jan 01 00:00:00 1970 +0000 summary: mergeCD +Check that the phase cache is properly invalidated after a strip with bookmark. + + $ cat > ../stripstalephasecache.py << EOF + > from mercurial import extensions, localrepo + > def transactioncallback(orig, repo, desc, *args, **kwargs): + > def test(transaction): + > # observe cache inconsistency + > try: + > [repo.changelog.node(r) for r in repo.revs("not public()")] + > except IndexError: + > repo.ui.status("Index error!\n") + > transaction = orig(repo, desc, *args, **kwargs) + > # warm up the phase cache + > list(repo.revs("not public()")) + > if desc != 'strip': + > transaction.addpostclose("phase invalidation test", test) + > return transaction + > def extsetup(ui): + > extensions.wrapfunction(localrepo.localrepository, "transaction", + > transactioncallback) + > EOF + $ hg up -C 2 + 1 files updated, 0 files merged, 0 files removed, 0 files unresolved + $ echo k > k + $ hg add k + $ hg commit -m commitK + $ echo l > l + $ hg add l + $ hg commit -m commitL + $ hg book -r tip blah + $ hg strip ".^" --config extensions.crash=$TESTTMP/stripstalephasecache.py + 0 files updated, 0 files merged, 2 files removed, 0 files unresolved + saved backup bundle to $TESTTMP/issue4736/.hg/strip-backup/8f0b4384875c-4fa10deb-backup.hg (glob) + $ hg up -C 1 + 0 files updated, 0 files merged, 1 files removed, 0 files unresolved Error during post-close callback of the strip transaction (They should be gracefully handled and reported)
# HG changeset patch # User Laurent Charignon <lcharignon@fb.com> # Date 1463058839 25200 # Thu May 12 06:13:59 2016 -0700 # Node ID 05aaad2606ba0f2c6a0bf088d528dacf1436c419 # Parent df838803c1d487e4601f96c6cfd85e6ad4f6291f strip: invalidate phase cache after stripping changeset (issue5235) When we remove a changeset from the changelog, the phase cache must be invalidated, otherwise it could refer to changesets that are no longer in the repo. To reproduce the failure, I created an extension querying the phase cache after the strip transaction is over. To do that, I stripped two commits with a bookmark on one of them to force another transaction (we open a transaction for moving bookmarks) after the strip transaction. Without the fix in this patch, the test leads to a stacktrace showing the issue: repair.strip(ui, repo, revs, backup) File "/Users/lcharignon/facebook-hg-rpms/hg-crew/mercurial/repair.py", line 205, in strip tr.close() File "/Users/lcharignon/facebook-hg-rpms/hg-crew/mercurial/transaction.py", line 44, in _active return func(self, *args, **kwds) File "/Users/lcharignon/facebook-hg-rpms/hg-crew/mercurial/transaction.py", line 490, in close self._postclosecallback[cat](self) File "$TESTTMP/crashstrip2.py", line 4, in test [repo.changelog.node(r) for r in repo.revs("not public()")] File "/Users/lcharignon/facebook-hg-rpms/hg-crew/mercurial/changelog.py", line 337, in node return super(changelog, self).node(rev) File "/Users/lcharignon/facebook-hg-rpms/hg-crew/mercurial/revlog.py", line 377, in node return self.index[rev][7] IndexError: revlog index out of range The situation was encountered in inhibit (evolve's repo) where we would crash following the volatile set invalidation submitted by Augie in e6f490e328635312ee214a12bc7fd3c7d46bf9ce. Before his patch the issue was masked as we were not accessing the phasecache after stripping a revision. This bug uncovered another but in histedit (see explanation in issue5235). I changed the histedit test accordingly to avoid fixing two things at once.