Mercurial > hg
view tests/test-rebase-interruptions.t @ 46607:e9901d01d135
revlog: add a mechanism to verify expected file position before appending
If someone uses `hg debuglocks`, or some non-hg process writes to the .hg
directory without respecting the locks, or if the repo's on a networked
filesystem, it's possible for the revlog code to write out corrupted data.
The form of this corruption can vary depending on what data was written and how
that happened. We are in the "networked filesystem" case (though I've had users
also do this to themselves with the "`hg debuglocks`" scenario), and most often
see this with the changelog. What ends up happening is we produce two items
(let's call them rev1 and rev2) in the .i file that have the same linkrev,
baserev, and offset into the .d file, while the data in the .d file is appended
properly. rev2's compressed_size is accurate for rev2, but when we go to
decompress the data in the .d file, we use the offset that's recorded in the
index file, which is the same as rev1, and attempt to decompress
rev2.compressed_size bytes of rev1's data. This usually does not succeed. :)
When using inline data, this also fails, though I haven't investigated why too
closely. This shows up as a "patch decode" error. I believe what's happening
there is that we're basically ignoring the offset field, getting the data
properly, but since baserev != rev, it thinks this is a delta based on rev
(instead of a full text) and can't actually apply it as such.
For now, I'm going to make this an optional component and default it to entirely
off. I may increase the default severity of this in the future, once I've
enabled it for my users and we gain more experience with it. Luckily, most of my
users have a versioned filesystem and can roll back to before the corruption has
been written, it's just a hassle to do so and not everyone knows how (so it's a
support burden). Users on other filesystems will not have that luxury, and this
can cause them to have a corrupted repository that they are unlikely to know how
to resolve, and they'll see this as a data-loss event. Refusing to create the
corruption is a much better user experience.
This mechanism is not perfect. There may be false-negatives (racy writes that
are not detected). There should not be any false-positives (non-racy writes that
are detected as such). This is not a mechanism that makes putting a repo on a
networked filesystem "safe" or "supported", just *less* likely to cause
corruption.
Differential Revision: https://phab.mercurial-scm.org/D9952
author | Kyle Lippincott <spectral@google.com> |
---|---|
date | Wed, 03 Feb 2021 16:33:10 -0800 |
parents | 768056549737 |
children |
line wrap: on
line source
$ cat >> $HGRCPATH <<EOF > [extensions] > rebase= > > [phases] > publish=False > > [alias] > tglog = log -G --template "{rev}: {node|short} '{desc}' {branches}\n" > tglogp = log -G --template "{rev}: {node|short} {phase} '{desc}' {branches}\n" > EOF $ hg init a $ cd a $ echo A > A $ hg ci -Am A adding A $ echo B > B $ hg ci -Am B adding B $ echo C >> A $ hg ci -m C $ hg up -q -C 0 $ echo D >> A $ hg ci -m D created new head $ echo E > E $ hg ci -Am E adding E $ cd .. Changes during an interruption - continue: $ hg clone -q -u . a a1 $ cd a1 $ hg tglog @ 4: ae36e8e3dfd7 'E' | o 3: 46b37eabc604 'D' | | o 2: 965c486023db 'C' | | | o 1: 27547f69f254 'B' |/ o 0: 4a2df7238c3b 'A' Rebasing B onto E: $ hg rebase -s 1 -d 4 rebasing 1:27547f69f254 "B" rebasing 2:965c486023db "C" merging A warning: conflicts while merging A! (edit, then use 'hg resolve --mark') unresolved conflicts (see 'hg resolve', then 'hg rebase --continue') [240] Force a commit on C during the interruption: $ hg up -q -C 2 --config 'extensions.rebase=!' $ echo 'Extra' > Extra $ hg add Extra $ hg ci -m 'Extra' --config 'extensions.rebase=!' Force this commit onto secret phase $ hg phase --force --secret 6 $ hg tglogp @ 6: deb5d2f93d8b secret 'Extra' | | o 5: 45396c49d53b draft 'B' | | | o 4: ae36e8e3dfd7 draft 'E' | | | o 3: 46b37eabc604 draft 'D' | | o | 2: 965c486023db draft 'C' | | o | 1: 27547f69f254 draft 'B' |/ o 0: 4a2df7238c3b draft 'A' Resume the rebasing: $ hg rebase --continue already rebased 1:27547f69f254 "B" as 45396c49d53b rebasing 2:965c486023db "C" merging A warning: conflicts while merging A! (edit, then use 'hg resolve --mark') unresolved conflicts (see 'hg resolve', then 'hg rebase --continue') [240] Solve the conflict and go on: $ echo 'conflict solved' > A $ rm A.orig $ hg resolve -m A (no more unresolved files) continue: hg rebase --continue $ hg rebase --continue already rebased 1:27547f69f254 "B" as 45396c49d53b rebasing 2:965c486023db "C" warning: orphaned descendants detected, not stripping 27547f69f254, 965c486023db $ hg tglogp o 7: d2d25e26288e draft 'C' | | o 6: deb5d2f93d8b secret 'Extra' | | o | 5: 45396c49d53b draft 'B' | | @ | 4: ae36e8e3dfd7 draft 'E' | | o | 3: 46b37eabc604 draft 'D' | | | o 2: 965c486023db draft 'C' | | | o 1: 27547f69f254 draft 'B' |/ o 0: 4a2df7238c3b draft 'A' $ cd .. Changes during an interruption - abort: $ hg clone -q -u . a a2 $ cd a2 $ hg tglog @ 4: ae36e8e3dfd7 'E' | o 3: 46b37eabc604 'D' | | o 2: 965c486023db 'C' | | | o 1: 27547f69f254 'B' |/ o 0: 4a2df7238c3b 'A' Rebasing B onto E: $ hg rebase -s 1 -d 4 rebasing 1:27547f69f254 "B" rebasing 2:965c486023db "C" merging A warning: conflicts while merging A! (edit, then use 'hg resolve --mark') unresolved conflicts (see 'hg resolve', then 'hg rebase --continue') [240] Force a commit on B' during the interruption: $ hg up -q -C 5 --config 'extensions.rebase=!' $ echo 'Extra' > Extra $ hg add Extra $ hg ci -m 'Extra' --config 'extensions.rebase=!' $ hg tglog @ 6: 402ee3642b59 'Extra' | o 5: 45396c49d53b 'B' | o 4: ae36e8e3dfd7 'E' | o 3: 46b37eabc604 'D' | | o 2: 965c486023db 'C' | | | o 1: 27547f69f254 'B' |/ o 0: 4a2df7238c3b 'A' Abort the rebasing: $ hg rebase --abort warning: new changesets detected on destination branch, can't strip rebase aborted $ hg tglog @ 6: 402ee3642b59 'Extra' | o 5: 45396c49d53b 'B' | o 4: ae36e8e3dfd7 'E' | o 3: 46b37eabc604 'D' | | o 2: 965c486023db 'C' | | | o 1: 27547f69f254 'B' |/ o 0: 4a2df7238c3b 'A' $ cd .. Changes during an interruption - abort (again): $ hg clone -q -u . a a3 $ cd a3 $ hg tglogp @ 4: ae36e8e3dfd7 draft 'E' | o 3: 46b37eabc604 draft 'D' | | o 2: 965c486023db draft 'C' | | | o 1: 27547f69f254 draft 'B' |/ o 0: 4a2df7238c3b draft 'A' Rebasing B onto E: $ hg rebase -s 1 -d 4 rebasing 1:27547f69f254 "B" rebasing 2:965c486023db "C" merging A warning: conflicts while merging A! (edit, then use 'hg resolve --mark') unresolved conflicts (see 'hg resolve', then 'hg rebase --continue') [240] Change phase on B and B' $ hg up -q -C 5 --config 'extensions.rebase=!' $ hg phase --public 1 $ hg phase --public 5 $ hg phase --secret -f 2 $ hg tglogp @ 5: 45396c49d53b public 'B' | o 4: ae36e8e3dfd7 public 'E' | o 3: 46b37eabc604 public 'D' | | o 2: 965c486023db secret 'C' | | | o 1: 27547f69f254 public 'B' |/ o 0: 4a2df7238c3b public 'A' Abort the rebasing: $ hg rebase --abort warning: can't clean up public changesets 45396c49d53b rebase aborted $ hg tglogp @ 5: 45396c49d53b public 'B' | o 4: ae36e8e3dfd7 public 'E' | o 3: 46b37eabc604 public 'D' | | o 2: 965c486023db secret 'C' | | | o 1: 27547f69f254 public 'B' |/ o 0: 4a2df7238c3b public 'A' Test rebase interrupted by hooks $ hg up 2 1 files updated, 0 files merged, 1 files removed, 0 files unresolved $ echo F > F $ hg add F $ hg ci -m F $ cd .. Continue rebase after upgrading from an hg version before 9c9cfecd4600: $ hg clone -q -u . a a4 $ cd a4 $ hg tglog @ 4: ae36e8e3dfd7 'E' | o 3: 46b37eabc604 'D' | | o 2: 965c486023db 'C' | | | o 1: 27547f69f254 'B' |/ o 0: 4a2df7238c3b 'A' $ hg rebase -s 1 -d 4 rebasing 1:27547f69f254 "B" rebasing 2:965c486023db "C" merging A warning: conflicts while merging A! (edit, then use 'hg resolve --mark') unresolved conflicts (see 'hg resolve', then 'hg rebase --continue') [240] $ hg tglog @ 5: 45396c49d53b 'B' | o 4: ae36e8e3dfd7 'E' | o 3: 46b37eabc604 'D' | | % 2: 965c486023db 'C' | | | o 1: 27547f69f254 'B' |/ o 0: 4a2df7238c3b 'A' Simulate having run the above with an older hg version by manually setting two dirstate parents. We should not get a merge commit when we continue. $ hg debugsetparents 5 2 $ echo 'conflict solved' > A $ hg resolve -m A (no more unresolved files) continue: hg rebase --continue $ hg rebase --continue already rebased 1:27547f69f254 "B" as 45396c49d53b rebasing 2:965c486023db "C" saved backup bundle to $TESTTMP/a4/.hg/strip-backup/27547f69f254-359abdd7-rebase.hg $ hg tglog o 4: d2d25e26288e 'C' | o 3: 45396c49d53b 'B' | @ 2: ae36e8e3dfd7 'E' | o 1: 46b37eabc604 'D' | o 0: 4a2df7238c3b 'A' $ cd .. (precommit version) $ cp -R a3 hook-precommit $ cd hook-precommit $ hg rebase --source 2 --dest 5 --tool internal:other --config 'hooks.precommit=hg status | grep "M A"' rebasing 2:965c486023db "C" M A rebasing 6:a0b2430ebfb8 tip "F" abort: precommit hook exited with status 1 [40] $ hg tglogp @ 7: 401ccec5e39f secret 'C' | | o 6: a0b2430ebfb8 secret 'F' | | o | 5: 45396c49d53b public 'B' | | o | 4: ae36e8e3dfd7 public 'E' | | o | 3: 46b37eabc604 public 'D' | | | o 2: 965c486023db secret 'C' | | | o 1: 27547f69f254 public 'B' |/ o 0: 4a2df7238c3b public 'A' $ hg rebase --continue already rebased 2:965c486023db "C" as 401ccec5e39f rebasing 6:a0b2430ebfb8 "F" saved backup bundle to $TESTTMP/hook-precommit/.hg/strip-backup/965c486023db-aa6250e7-rebase.hg $ hg tglogp @ 6: 6e92a149ac6b secret 'F' | o 5: 401ccec5e39f secret 'C' | o 4: 45396c49d53b public 'B' | o 3: ae36e8e3dfd7 public 'E' | o 2: 46b37eabc604 public 'D' | | o 1: 27547f69f254 public 'B' |/ o 0: 4a2df7238c3b public 'A' $ cd .. (pretxncommit version) $ cp -R a3 hook-pretxncommit $ cd hook-pretxncommit $ hg rebase --source 2 --dest 5 --tool internal:other \ > --config 'hooks.tonative.pretxncommit=True' --config 'hooks.pretxncommit=hg log -r $HG_NODE | grep "summary: C"' rebasing 2:965c486023db "C" summary: C rebasing 6:a0b2430ebfb8 tip "F" transaction abort! rollback completed abort: pretxncommit hook exited with status 1 [40] $ hg tglogp @ 7: 401ccec5e39f secret 'C' | | o 6: a0b2430ebfb8 secret 'F' | | o | 5: 45396c49d53b public 'B' | | o | 4: ae36e8e3dfd7 public 'E' | | o | 3: 46b37eabc604 public 'D' | | | o 2: 965c486023db secret 'C' | | | o 1: 27547f69f254 public 'B' |/ o 0: 4a2df7238c3b public 'A' $ hg rebase --continue already rebased 2:965c486023db "C" as 401ccec5e39f rebasing 6:a0b2430ebfb8 "F" saved backup bundle to $TESTTMP/hook-pretxncommit/.hg/strip-backup/965c486023db-aa6250e7-rebase.hg $ hg tglogp @ 6: 6e92a149ac6b secret 'F' | o 5: 401ccec5e39f secret 'C' | o 4: 45396c49d53b public 'B' | o 3: ae36e8e3dfd7 public 'E' | o 2: 46b37eabc604 public 'D' | | o 1: 27547f69f254 public 'B' |/ o 0: 4a2df7238c3b public 'A' $ cd .. (pretxnclose version) $ cp -R a3 hook-pretxnclose $ cd hook-pretxnclose $ hg rebase --source 2 --dest 5 --tool internal:other --config 'hooks.pretxnclose=hg log -r tip | grep "summary: C"' rebasing 2:965c486023db "C" summary: C rebasing 6:a0b2430ebfb8 tip "F" transaction abort! rollback completed abort: pretxnclose hook exited with status 1 [40] $ hg tglogp @ 7: 401ccec5e39f secret 'C' | | o 6: a0b2430ebfb8 secret 'F' | | o | 5: 45396c49d53b public 'B' | | o | 4: ae36e8e3dfd7 public 'E' | | o | 3: 46b37eabc604 public 'D' | | | o 2: 965c486023db secret 'C' | | | o 1: 27547f69f254 public 'B' |/ o 0: 4a2df7238c3b public 'A' $ hg rebase --continue already rebased 2:965c486023db "C" as 401ccec5e39f rebasing 6:a0b2430ebfb8 "F" saved backup bundle to $TESTTMP/hook-pretxnclose/.hg/strip-backup/965c486023db-aa6250e7-rebase.hg $ hg tglogp @ 6: 6e92a149ac6b secret 'F' | o 5: 401ccec5e39f secret 'C' | o 4: 45396c49d53b public 'B' | o 3: ae36e8e3dfd7 public 'E' | o 2: 46b37eabc604 public 'D' | | o 1: 27547f69f254 public 'B' |/ o 0: 4a2df7238c3b public 'A' $ cd .. Make sure merge state is cleaned up after a no-op rebase merge (issue5494) $ hg init repo $ cd repo $ echo a > a $ hg commit -qAm base $ echo b >> a $ hg commit -qm b $ hg up '.^' 1 files updated, 0 files merged, 0 files removed, 0 files unresolved $ echo c >> a $ hg commit -qm c $ hg rebase -s 1 -d 2 --noninteractive rebasing 1:fdaca8533b86 "b" merging a warning: conflicts while merging a! (edit, then use 'hg resolve --mark') unresolved conflicts (see 'hg resolve', then 'hg rebase --continue') [240] $ echo a > a $ echo c >> a $ hg resolve --mark a (no more unresolved files) continue: hg rebase --continue $ hg rebase --continue rebasing 1:fdaca8533b86 "b" note: not rebasing 1:fdaca8533b86 "b", its destination already has all its changes saved backup bundle to $TESTTMP/repo/.hg/strip-backup/fdaca8533b86-7fd70513-rebase.hg $ hg resolve --list $ test -d .hg/merge [1] Now try again with --collapse $ hg unbundle -q .hg/strip-backup/fdaca8533b86-7fd70513-rebase.hg $ hg rebase -s 2 -d 1 --noninteractive --collapse rebasing 2:fdaca8533b86 tip "b" merging a warning: conflicts while merging a! (edit, then use 'hg resolve --mark') unresolved conflicts (see 'hg resolve', then 'hg rebase --continue') [240] $ echo a > a $ echo c >> a $ hg resolve --mark a (no more unresolved files) continue: hg rebase --continue $ hg rebase --continue rebasing 2:fdaca8533b86 tip "b" saved backup bundle to $TESTTMP/repo/.hg/strip-backup/fdaca8533b86-7fd70513-rebase.hg $ hg resolve --list $ test -d .hg/merge [1]