Mercurial > hg
view tests/test-copy-move-merge.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 | 69429972ff1f |
children | 38941a28406a |
line wrap: on
line source
Test for the full copytracing algorithm ======================================= Initial Setup ============= use git diff to see rename $ cat << EOF >> $HGRCPATH > [diff] > git=yes > EOF Setup an history where one side copy and rename a file (and update it) while the other side update it. $ hg init t $ cd t $ echo 1 > a $ hg ci -qAm "first" $ hg cp a b $ hg mv a c $ echo 2 >> b $ echo 2 >> c $ hg ci -qAm "second" $ hg co -C 0 1 files updated, 0 files merged, 2 files removed, 0 files unresolved $ echo 0 > a $ echo 1 >> a $ hg ci -qAm "other" $ hg log -G --patch @ changeset: 2:add3f11052fa | tag: tip | parent: 0:b8bf91eeebbc | user: test | date: Thu Jan 01 00:00:00 1970 +0000 | summary: other | | diff --git a/a b/a | --- a/a | +++ b/a | @@ -1,1 +1,2 @@ | +0 | 1 | | o changeset: 1:17c05bb7fcb6 |/ user: test | date: Thu Jan 01 00:00:00 1970 +0000 | summary: second | | diff --git a/a b/b | rename from a | rename to b | --- a/a | +++ b/b | @@ -1,1 +1,2 @@ | 1 | +2 | diff --git a/a b/c | copy from a | copy to c | --- a/a | +++ b/c | @@ -1,1 +1,2 @@ | 1 | +2 | o changeset: 0:b8bf91eeebbc user: test date: Thu Jan 01 00:00:00 1970 +0000 summary: first diff --git a/a b/a new file mode 100644 --- /dev/null +++ b/a @@ -0,0 +1,1 @@ +1 Test Simple Merge ================= $ hg merge --debug unmatched files in other: b c all copies found (* = to merge, ! = divergent, % = renamed and deleted): on remote side: src: 'a' -> dst: 'b' * src: 'a' -> dst: 'c' * checking for directory renames resolving manifests branchmerge: True, force: False, partial: False ancestor: b8bf91eeebbc, local: add3f11052fa+, remote: 17c05bb7fcb6 starting 4 threads for background file closing (?) preserving a for resolve of b preserving a for resolve of c removing a b: remote moved from a -> m (premerge) picked tool ':merge' for b (binary False symlink False changedelete False) merging a and b to b my b@add3f11052fa+ other b@17c05bb7fcb6 ancestor a@b8bf91eeebbc premerge successful c: remote moved from a -> m (premerge) picked tool ':merge' for c (binary False symlink False changedelete False) merging a and c to c my c@add3f11052fa+ other c@17c05bb7fcb6 ancestor a@b8bf91eeebbc premerge successful 0 files updated, 2 files merged, 0 files removed, 0 files unresolved (branch merge, don't forget to commit) file b $ cat b 0 1 2 file c $ cat c 0 1 2 Test disabling copy tracing =========================== first verify copy metadata was kept ----------------------------------- $ hg up -qC 2 $ hg rebase --keep -d 1 -b 2 --config extensions.rebase= rebasing 2:add3f11052fa tip "other" merging b and a to b merging c and a to c $ cat b 0 1 2 next verify copy metadata is lost when disabled ------------------------------------------------ $ hg strip -r . --config extensions.strip= 2 files updated, 0 files merged, 0 files removed, 0 files unresolved saved backup bundle to $TESTTMP/t/.hg/strip-backup/550bd84c0cd3-fc575957-backup.hg $ hg up -qC 2 $ hg rebase --keep -d 1 -b 2 --config extensions.rebase= --config experimental.copytrace=off --config ui.interactive=True << EOF > c > EOF rebasing 2:add3f11052fa tip "other" file 'a' was deleted in local [dest] but was modified in other [source]. You can use (c)hanged version, leave (d)eleted, or leave (u)nresolved. What do you want to do? c $ cat b 1 2 $ cd .. Verify disabling copy tracing still keeps copies from rebase source ------------------------------------------------------------------- $ hg init copydisable $ cd copydisable $ touch a $ hg ci -Aqm 'add a' $ touch b $ hg ci -Aqm 'add b, c' $ hg cp b x $ echo x >> x $ hg ci -qm 'copy b->x' $ hg up -q 1 $ touch z $ hg ci -Aqm 'add z' $ hg log -G -T '{rev} {desc}\n' @ 3 add z | | o 2 copy b->x |/ o 1 add b, c | o 0 add a $ hg rebase -d . -b 2 --config extensions.rebase= --config experimental.copytrace=off rebasing 2:6adcf8c12e7d "copy b->x" saved backup bundle to $TESTTMP/copydisable/.hg/strip-backup/6adcf8c12e7d-ce4b3e75-rebase.hg $ hg up -q 3 $ hg log -f x -T '{rev} {desc}\n' 3 copy b->x 1 add b, c $ cd ../ test storage preservation ------------------------- Verify rebase do not discard recorded copies data when copy tracing usage is disabled. Setup $ hg init copydisable3 $ cd copydisable3 $ touch a $ hg ci -Aqm 'add a' $ hg cp a b $ hg ci -Aqm 'copy a->b' $ hg mv b c $ hg ci -Aqm 'move b->c' $ hg up -q 0 $ hg cp a b $ echo b >> b $ hg ci -Aqm 'copy a->b (2)' $ hg log -G -T '{rev} {desc}\n' @ 3 copy a->b (2) | | o 2 move b->c | | | o 1 copy a->b |/ o 0 add a Actual Test A file is copied on one side and has been moved twice on the other side. the file is copied from `0:a`, so the file history of the `3:b` should trace directly to `0:a`. $ hg rebase -d 2 -s 3 --config extensions.rebase= --config experimental.copytrace=off rebasing 3:47e1a9e6273b tip "copy a->b (2)" saved backup bundle to $TESTTMP/copydisable3/.hg/strip-backup/47e1a9e6273b-2d099c59-rebase.hg $ hg log -G -f b @ changeset: 3:76024fb4b05b : tag: tip : user: test : date: Thu Jan 01 00:00:00 1970 +0000 : summary: copy a->b (2) : o changeset: 0:ac82d8b1f7c4 user: test date: Thu Jan 01 00:00:00 1970 +0000 summary: add a