Mercurial > hg
view tests/test-copies-unrelated.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 | 16c361152133 |
children |
line wrap: on
line source
#testcases filelog compatibility changeset sidedata $ cat >> $HGRCPATH << EOF > [extensions] > rebase= > [alias] > l = log -G -T '{rev} {desc}\n{files}\n' > EOF #if compatibility $ cat >> $HGRCPATH << EOF > [experimental] > copies.read-from = compatibility > EOF #endif #if changeset $ cat >> $HGRCPATH << EOF > [experimental] > copies.read-from = changeset-only > copies.write-to = changeset-only > EOF #endif #if sidedata $ cat >> $HGRCPATH << EOF > [format] > exp-use-copies-side-data-changeset = yes > EOF #endif $ REPONUM=0 $ newrepo() { > cd $TESTTMP > REPONUM=`expr $REPONUM + 1` > hg init repo-$REPONUM > cd repo-$REPONUM > } Copy a file, then delete destination, then copy again. This does not create a new filelog entry. $ newrepo $ echo x > x $ hg ci -Aqm 'add x' $ echo x2 > x $ hg ci -m 'modify x' $ hg co -q 0 $ hg cp x y $ hg ci -qm 'copy x to y' $ hg rm y $ hg ci -m 'remove y' $ hg cp -f x y $ hg ci -m 'copy x onto y (again)' $ hg l @ 4 copy x onto y (again) | y o 3 remove y | y o 2 copy x to y | y | o 1 modify x |/ x o 0 add x x $ hg debugp1copies -r 4 x -> y $ hg debugpathcopies 0 4 x -> y $ hg graft -r 1 grafting 1:* "modify x" (glob) merging y and x to y $ hg co -qC 1 $ hg graft -r 4 grafting 4:* "copy x onto y (again)" (glob) merging x and y to y Copy x to y, then remove y, then add back y. With copy metadata in the changeset, this could easily end up reporting y as copied from x (if we don't unmark it as a copy when it's removed). Despite x and y not being related, we want grafts to propagate across the rename. $ newrepo $ echo x > x $ hg ci -Aqm 'add x' $ echo x2 > x $ hg ci -m 'modify x' $ hg co -q 0 $ hg mv x y $ hg ci -qm 'rename x to y' $ hg rm y $ hg ci -qm 'remove y' $ echo x > y $ hg ci -Aqm 'add back y' $ hg l @ 4 add back y | y o 3 remove y | y o 2 rename x to y | x y | o 1 modify x |/ x o 0 add x x $ hg debugpathcopies 0 4 BROKEN: This should succeed and merge the changes from x into y $ hg graft -r 1 grafting 1:* "modify x" (glob) file 'x' was deleted in local [local] but was modified in other [graft]. You can use (c)hanged version, leave (d)eleted, or leave (u)nresolved. What do you want to do? u abort: unresolved conflicts, can't continue (use 'hg resolve' and 'hg graft --continue') [1] Add x, remove it, then add it back, then rename x to y. Similar to the case above, but here the break in history is before the rename. $ newrepo $ echo x > x $ hg ci -Aqm 'add x' $ echo x2 > x $ hg ci -m 'modify x' $ hg co -q 0 $ hg rm x $ hg ci -qm 'remove x' $ echo x > x $ hg ci -Aqm 'add x again' $ hg mv x y $ hg ci -m 'rename x to y' $ hg l @ 4 rename x to y | x y o 3 add x again | x o 2 remove x | x | o 1 modify x |/ x o 0 add x x $ hg debugpathcopies 0 4 x -> y $ hg graft -r 1 grafting 1:* "modify x" (glob) merging y and x to y $ hg co -qC 1 $ hg graft -r 4 grafting 4:* "rename x to y" (glob) merging x and y to y Add x, modify it, remove it, then add it back, then rename x to y. Similar to the case above, but here the re-added file's nodeid is different from before the break. $ newrepo $ echo x > x $ hg ci -Aqm 'add x' $ echo x2 > x $ hg ci -m 'modify x' $ echo x3 > x $ hg ci -qm 'modify x again' $ hg co -q 1 $ hg rm x $ hg ci -qm 'remove x' # Same content to avoid conflicts $ hg revert -r 1 x $ hg ci -Aqm 'add x again' $ hg mv x y $ hg ci -m 'rename x to y' $ hg l @ 5 rename x to y | x y o 4 add x again | x o 3 remove x | x | o 2 modify x again |/ x o 1 modify x | x o 0 add x x $ hg debugpathcopies 0 5 x -> y (no-filelog !) #if no-filelog $ hg graft -r 2 grafting 2:* "modify x again" (glob) merging y and x to y #else BROKEN: This should succeed and merge the changes from x into y $ hg graft -r 2 grafting 2:* "modify x again" (glob) file 'x' was deleted in local [local] but was modified in other [graft]. You can use (c)hanged version, leave (d)eleted, or leave (u)nresolved. What do you want to do? u abort: unresolved conflicts, can't continue (use 'hg resolve' and 'hg graft --continue') [1] #endif $ hg co -qC 2 BROKEN: This should succeed and merge the changes from x into y $ hg graft -r 5 grafting 5:* "rename x to y"* (glob) file 'x' was deleted in other [graft] but was modified in local [local]. You can use (c)hanged version, (d)elete, or leave (u)nresolved. What do you want to do? u abort: unresolved conflicts, can't continue (use 'hg resolve' and 'hg graft --continue') [1] Add x, remove it, then add it back, rename x to y from the first commit. Similar to the case above, but here the break in history is parallel to the rename. $ newrepo $ echo x > x $ hg ci -Aqm 'add x' $ hg rm x $ hg ci -qm 'remove x' $ echo x > x $ hg ci -Aqm 'add x again' $ echo x2 > x $ hg ci -m 'modify x' $ hg co -q 0 $ hg mv x y $ hg ci -qm 'rename x to y' $ hg l @ 4 rename x to y | x y | o 3 modify x | | x | o 2 add x again | | x | o 1 remove x |/ x o 0 add x x $ hg debugpathcopies 2 4 x -> y $ hg graft -r 3 grafting 3:* "modify x" (glob) merging y and x to y $ hg co -qC 3 $ hg graft -r 4 grafting 4:* "rename x to y" (glob) merging x and y to y Add x, remove it, then add it back, rename x to y from the first commit. Similar to the case above, but here the re-added file's nodeid is different from the base. $ newrepo $ echo x > x $ hg ci -Aqm 'add x' $ hg rm x $ hg ci -qm 'remove x' $ echo x2 > x $ hg ci -Aqm 'add x again with different content' $ hg co -q 0 $ hg mv x y $ hg ci -qm 'rename x to y' $ hg l @ 3 rename x to y | x y | o 2 add x again with different content | | x | o 1 remove x |/ x o 0 add x x $ hg debugpathcopies 2 3 x -> y BROKEN: This should merge the changes from x into y $ hg graft -r 2 grafting 2:* "add x again with different content" (glob) $ hg co -qC 2 BROKEN: This should succeed and merge the changes from x into y $ hg graft -r 3 grafting 3:* "rename x to y" (glob) file 'x' was deleted in other [graft] but was modified in local [local]. You can use (c)hanged version, (d)elete, or leave (u)nresolved. What do you want to do? u abort: unresolved conflicts, can't continue (use 'hg resolve' and 'hg graft --continue') [1] Add x on two branches, then rename x to y on one side. Similar to the case above, but here the break in history is via the base commit. $ newrepo $ echo a > a $ hg ci -Aqm 'base' $ echo x > x $ hg ci -Aqm 'add x' $ echo x2 > x $ hg ci -m 'modify x' $ hg co -q 0 $ echo x > x $ hg ci -Aqm 'add x again' $ hg mv x y $ hg ci -qm 'rename x to y' $ hg l @ 4 rename x to y | x y o 3 add x again | x | o 2 modify x | | x | o 1 add x |/ x o 0 base a $ hg debugpathcopies 1 4 x -> y $ hg graft -r 2 grafting 2:* "modify x" (glob) merging y and x to y $ hg co -qC 2 $ hg graft -r 4 grafting 4:* "rename x to y"* (glob) merging x and y to y Add x on two branches, with same content but different history, then rename x to y on one side. Similar to the case above, here the file's nodeid is different between the branches. $ newrepo $ echo a > a $ hg ci -Aqm 'base' $ echo x > x $ hg ci -Aqm 'add x' $ echo x2 > x $ hg ci -m 'modify x' $ hg co -q 0 $ touch x $ hg ci -Aqm 'add empty x' # Same content to avoid conflicts $ hg revert -r 1 x $ hg ci -m 'modify x to match commit 1' $ hg mv x y $ hg ci -qm 'rename x to y' $ hg l @ 5 rename x to y | x y o 4 modify x to match commit 1 | x o 3 add empty x | x | o 2 modify x | | x | o 1 add x |/ x o 0 base a $ hg debugpathcopies 1 5 x -> y (no-filelog !) #if no-filelog $ hg graft -r 2 grafting 2:* "modify x" (glob) merging y and x to y #else BROKEN: This should succeed and merge the changes from x into y $ hg graft -r 2 grafting 2:* "modify x" (glob) file 'x' was deleted in local [local] but was modified in other [graft]. You can use (c)hanged version, leave (d)eleted, or leave (u)nresolved. What do you want to do? u abort: unresolved conflicts, can't continue (use 'hg resolve' and 'hg graft --continue') [1] #endif $ hg co -qC 2 BROKEN: This should succeed and merge the changes from x into y $ hg graft -r 5 grafting 5:* "rename x to y"* (glob) file 'x' was deleted in other [graft] but was modified in local [local]. You can use (c)hanged version, (d)elete, or leave (u)nresolved. What do you want to do? u abort: unresolved conflicts, can't continue (use 'hg resolve' and 'hg graft --continue') [1] Copies via null revision (there shouldn't be any) $ newrepo $ echo x > x $ hg ci -Aqm 'add x' $ hg cp x y $ hg ci -m 'copy x to y' $ hg co -q null $ echo x > x $ hg ci -Aqm 'add x (again)' $ hg l @ 2 add x (again) x o 1 copy x to y | y o 0 add x x $ hg debugpathcopies 1 2 $ hg debugpathcopies 2 1 $ hg graft -r 1 grafting 1:* "copy x to y" (glob) Copies involving a merge of multiple roots. $ newrepo $ echo a > a $ hg ci -Aqm 'add a' $ echo a >> a $ hg ci -Aqm 'update a' $ echo a >> a $ hg ci -Aqm 'update a' $ hg up null 0 files updated, 0 files merged, 1 files removed, 0 files unresolved $ echo b > a $ hg ci -Aqm 'add a' $ hg mv a b $ hg ci -Aqm 'move a to b' $ echo b >> b $ hg ci -Aqm 'update b' $ hg merge 0 1 files updated, 0 files merged, 0 files removed, 0 files unresolved (branch merge, don't forget to commit) $ hg ci -m "merge with other branch" $ echo a >> a $ echo a >> a $ echo b >> b $ hg ci -Aqm 'update a and b' $ hg l @ 7 update a and b | a b o 6 merge with other branch |\ | o 5 update b | | b | o 4 move a to b | | a b | o 3 add a | a | o 2 update a | | a | o 1 update a |/ a o 0 add a a $ hg cat a -r 7 a a a $ hg cat a -r 2 a a a $ hg cat a -r 0 a $ hg debugpathcopies 7 2 $ hg debugpathcopies 2 7 $ hg merge 2 1 files updated, 0 files merged, 0 files removed, 0 files unresolved (branch merge, don't forget to commit)