Mercurial > hg
view tests/test-url-rev.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 | 4441705b7111 |
children |
line wrap: on
line source
Test basic functionality of url#rev syntax $ hg init repo $ cd repo $ echo a > a $ hg ci -qAm 'add a' $ hg branch foo marked working directory as branch foo (branches are permanent and global, did you want a bookmark?) $ echo >> a $ hg ci -m 'change a' $ cd .. $ hg clone 'repo#foo' clone adding changesets adding manifests adding file changes added 2 changesets with 2 changes to 1 files new changesets 1f0dee641bb7:cd2a86ecc814 updating to branch foo 1 files updated, 0 files merged, 0 files removed, 0 files unresolved $ hg --cwd clone heads changeset: 1:cd2a86ecc814 branch: foo tag: tip user: test date: Thu Jan 01 00:00:00 1970 +0000 summary: change a changeset: 0:1f0dee641bb7 user: test date: Thu Jan 01 00:00:00 1970 +0000 summary: add a $ hg --cwd clone parents changeset: 1:cd2a86ecc814 branch: foo tag: tip user: test date: Thu Jan 01 00:00:00 1970 +0000 summary: change a $ cat clone/.hg/hgrc # example repository config (see 'hg help config' for more info) [paths] default = $TESTTMP/repo#foo # path aliases to other clones of this repo in URLs or filesystem paths # (see 'hg help config.paths' for more info) # # default:pushurl = ssh://jdoe@example.net/hg/jdoes-fork # my-fork = ssh://jdoe@example.net/hg/jdoes-fork # my-clone = /home/jdoe/jdoes-clone [ui] # name and email (local to this repository, optional), e.g. # username = Jane Doe <jdoe@example.com> Changing original repo: $ cd repo $ echo >> a $ hg ci -m 'new head of branch foo' $ hg up -qC default $ echo bar > bar $ hg ci -qAm 'add bar' $ hg log changeset: 3:4cd725637392 tag: tip parent: 0:1f0dee641bb7 user: test date: Thu Jan 01 00:00:00 1970 +0000 summary: add bar changeset: 2:faba9097cad4 branch: foo user: test date: Thu Jan 01 00:00:00 1970 +0000 summary: new head of branch foo changeset: 1:cd2a86ecc814 branch: foo user: test date: Thu Jan 01 00:00:00 1970 +0000 summary: change a changeset: 0:1f0dee641bb7 user: test date: Thu Jan 01 00:00:00 1970 +0000 summary: add a $ hg -q outgoing '../clone' 2:faba9097cad4 3:4cd725637392 $ hg summary --remote --config paths.default='../clone' parent: 3:4cd725637392 tip add bar branch: default commit: (clean) update: (current) phases: 4 draft remote: 2 outgoing $ hg -q outgoing '../clone#foo' 2:faba9097cad4 $ hg summary --remote --config paths.default='../clone#foo' parent: 3:4cd725637392 tip add bar branch: default commit: (clean) update: (current) phases: 4 draft remote: 1 outgoing $ hg -q --cwd ../clone incoming '../repo#foo' 2:faba9097cad4 $ hg --cwd ../clone summary --remote --config paths.default='../repo#foo' parent: 1:cd2a86ecc814 tip change a branch: foo commit: (clean) update: (current) remote: 1 or more incoming $ hg -q push '../clone#foo' $ hg --cwd ../clone heads changeset: 2:faba9097cad4 branch: foo tag: tip user: test date: Thu Jan 01 00:00:00 1970 +0000 summary: new head of branch foo changeset: 0:1f0dee641bb7 user: test date: Thu Jan 01 00:00:00 1970 +0000 summary: add a $ hg -q --cwd ../clone incoming '../repo#foo' [1] $ hg --cwd ../clone summary --remote --config paths.default='../repo#foo' parent: 1:cd2a86ecc814 change a branch: foo commit: (clean) update: 1 new changesets (update) remote: (synced) $ cd .. $ cd clone $ hg rollback repository tip rolled back to revision 1 (undo push) $ hg -q incoming 2:faba9097cad4 $ hg -q pull $ hg heads changeset: 2:faba9097cad4 branch: foo tag: tip user: test date: Thu Jan 01 00:00:00 1970 +0000 summary: new head of branch foo changeset: 0:1f0dee641bb7 user: test date: Thu Jan 01 00:00:00 1970 +0000 summary: add a Pull should not have updated: $ hg parents -q 1:cd2a86ecc814 Going back to the default branch: $ hg up -C 0 1 files updated, 0 files merged, 0 files removed, 0 files unresolved $ hg parents changeset: 0:1f0dee641bb7 user: test date: Thu Jan 01 00:00:00 1970 +0000 summary: add a No new revs, no update: $ hg pull -qu $ hg parents -q 0:1f0dee641bb7 $ hg rollback repository tip rolled back to revision 1 (undo pull) $ hg parents -q 0:1f0dee641bb7 Pull -u takes us back to branch foo: $ hg pull -qu $ hg parents changeset: 2:faba9097cad4 branch: foo tag: tip user: test date: Thu Jan 01 00:00:00 1970 +0000 summary: new head of branch foo $ hg rollback repository tip rolled back to revision 1 (undo pull) working directory now based on revision 0 $ hg up -C 0 1 files updated, 0 files merged, 0 files removed, 0 files unresolved $ hg parents -q 0:1f0dee641bb7 $ hg heads -q 1:cd2a86ecc814 0:1f0dee641bb7 $ hg pull -qur default default $ hg parents changeset: 3:4cd725637392 tag: tip parent: 0:1f0dee641bb7 user: test date: Thu Jan 01 00:00:00 1970 +0000 summary: add bar $ hg heads changeset: 3:4cd725637392 tag: tip parent: 0:1f0dee641bb7 user: test date: Thu Jan 01 00:00:00 1970 +0000 summary: add bar changeset: 2:faba9097cad4 branch: foo user: test date: Thu Jan 01 00:00:00 1970 +0000 summary: new head of branch foo Test handling of invalid urls $ hg id http://foo/?bar abort: unsupported URL component: "bar" [255] $ cd .. Test handling common incoming revisions between "default" and "default-push" $ hg -R clone rollback repository tip rolled back to revision 1 (undo pull) working directory now based on revision 0 $ cd repo $ hg update -q -C default $ echo modified >> bar $ hg commit -m "new head to push current default head" $ hg -q push -r ".^1" '../clone' $ hg -q outgoing '../clone' 2:faba9097cad4 4:d515801a8f3d $ hg summary --remote --config paths.default='../clone#default' --config paths.default-push='../clone#foo' parent: 4:d515801a8f3d tip new head to push current default head branch: default commit: (clean) update: (current) phases: 1 draft remote: 1 outgoing $ hg summary --remote --config paths.default='../clone#foo' --config paths.default-push='../clone' parent: 4:d515801a8f3d tip new head to push current default head branch: default commit: (clean) update: (current) phases: 1 draft remote: 2 outgoing $ hg summary --remote --config paths.default='../clone' --config paths.default-push='../clone#foo' parent: 4:d515801a8f3d tip new head to push current default head branch: default commit: (clean) update: (current) phases: 1 draft remote: 1 outgoing $ hg clone -q -r 0 . ../another $ hg -q outgoing '../another#default' 3:4cd725637392 4:d515801a8f3d $ hg summary --remote --config paths.default='../another#default' --config paths.default-push='../clone#default' parent: 4:d515801a8f3d tip new head to push current default head branch: default commit: (clean) update: (current) phases: 1 draft remote: 1 outgoing $ cd .. Test url#rev syntax of local destination path, which should be taken as a 'url#rev' path $ hg clone repo '#foo' updating to branch default 2 files updated, 0 files merged, 0 files removed, 0 files unresolved $ hg root -R '#foo' $TESTTMP/#foo