Raphaël Gomès <rgomes@octobus.net> [Thu, 03 Nov 2022 15:44:54 +0100] rev 49563
config: fix indentation of some`share-safe` options
This makes the output much more readable.
Pierre-Yves David <pierre-yves.david@octobus.net> [Sun, 16 Oct 2022 04:48:21 +0200] rev 49562
perf-unbundle: do a quick and dirty fix to make it run on more commit
Without this change, the perf commands fails within the
f67741e8264b::
18415fc918a1 range (boundary excluded).
Check inline comment for details.
With this fix, the command is able to run on this range, with a slightly
different behavior (as no revset is "uninlined"). However this is still much
better than not being able to run anything in this range. Especially because
that range do see some performance regression for unbundle.
Pierre-Yves David <pierre-yves.david@octobus.net> [Wed, 19 Oct 2022 01:54:04 +0200] rev 49561
perf-unbundle: pre-indent the main block in per::unbundle
This makes the next changeset clearer.
Jason R. Coombs <jaraco@jaraco.com> [Wed, 19 Oct 2022 16:23:42 -0400] rev 49560
shelve: handle empty parents and nodestoremove in shelvedstate (
issue6748)
Raphaël Gomès <rgomes@octobus.net> [Thu, 03 Nov 2022 14:58:58 +0100] rev 49559
rhg: fallback when encountering ellipsis revisions
Ellipsis revisions are still experimental and buggy in non-trivial
histories. We currently have no plans to improve this situation nor
to add support for ellipsis revisions in `rhg`.
Falling back should be done carefully (since we may have already done
some work that is visible to the user), but in this case it's highly
unlikely that we're doing anything useful with a repo with ellipsis
revisions, so this should be strictly better, also since the
error message is more explicit.
Raphaël Gomès <rgomes@octobus.net> [Wed, 02 Nov 2022 12:05:34 +0100] rev 49558
dirstate-v2: hash the source of the ignore patterns as well
Fixes the test introduced in the last changeset. This caused the hash
to change, which means that the check in the test had to be adapted.
Since this hash is only done as a caching mechanism, invalidation does
not pose any backwards compatibility issues.
Raphaël Gomès <rgomes@octobus.net> [Wed, 02 Nov 2022 15:24:39 +0100] rev 49557
dirstate-v2: add test that shows a collision in ignore patterns hash
This hash is used for optimizing dirstate `status`. We demonstrate that the
hash is incorrectly ignoring the changes to the semantics of the ignore
files just because the contents (but not their source) haven't changed.
This is fixed in the next changeset.
Raphaël Gomès <rgomes@octobus.net> [Thu, 03 Nov 2022 12:08:02 +0100] rev 49556
testlib: add `--raw-sha1` option to `f`
This will be used in the patch fixing the ignore hash collision in
dirstate-v2 to check the behavior of the new hashing function.
Raphaël Gomès <rgomes@octobus.net> [Wed, 19 Oct 2022 16:28:41 +0200] rev 49555
rust-status: save new dircache even if just invalidated
There was a functional race between invalidating the cache (not acted upon
until the end of the status algorithm) and populating the new cache (which
relies upon an up-to-date version of the cache).
We simply inform the cache populating function that we've just invalidated
the cache for this particular directory since the information is present in
the same scope.
Raphaël Gomès <rgomes@octobus.net> [Wed, 19 Oct 2022 15:11:05 +0200] rev 49554
rust-status: fix typos and add docstrings to dircache related fields
Arseniy Alekseyev <aalekseyev@janestreet.com> [Thu, 22 Sep 2022 15:34:27 -0400] rev 49553
rhg: show a bug where repeated [hg status] is needed to cache everything
Pierre-Yves David <pierre-yves.david@octobus.net> [Fri, 04 Nov 2022 16:15:12 -0400] rev 49552
upgrade: no longer keep all revlogs in memory at any point
Keeping all object open is unsustainable, so we will open them on demand. This
mean opening them multiple times, but this is a lesser evil.
Each revlog consume a small amount of memory (index content, associated nodemap,
etc). While there are few "big" revlog, the sheer amount of small filelog can
become a significant issue memory wise, consuming multiple GB of memory. If you
combines this extra usage with the use of multiprocessing, this usage can
quickly get out of control. This can effectively block the upgrade of larger
repository. This changeset fixes this issue.
Matt Harbison <matt_harbison@yahoo.com> [Wed, 02 Nov 2022 14:23:09 -0400] rev 49551
demandimport: convert ignored modules from bytes -> str in extensions
The default list of ignored modules are str, and test for bypassing the lazy
import is `module.__name__ in ignores`, so these were effectively NOT ignored.
Most of these date back to the grand byteification in
687b865b95ad, with some
subsequent additions that followed the existing example.
I have no idea if these modules in fact need to bypass lazy importing, but at
least it follows the intent of the code.
Raphaël Gomès <rgomes@octobus.net> [Wed, 26 Oct 2022 18:46:56 +0200] rev 49550
dirstate-v2: fix edge case where entries aren't sorted
See previous commit for more details.
Raphaël Gomès <rgomes@octobus.net> [Wed, 26 Oct 2022 18:24:34 +0200] rev 49549
dirstate-v2: highlight a bug when Python-packed but used in `rhg`
The Python packer creates unsorted entries in the edge case that a file
starts with the same name as a sibling folder.
This bug has no effect on the Python `hg status` since Python ignores
directories. `rhg` assumes that all on-disk entries are sorted (which is
a property of the format) including folder, hence the issue highlighted.
This is also technically broken in Rust-augmented `hg status`, but it
makes setting up the test more complex than necessary, since it requires
the packing to be Python only (which it isn't if you have Rust extensions).
Fix is in the next commit.
Raphaël Gomès <rgomes@octobus.net> [Wed, 26 Oct 2022 12:20:23 +0200] rev 49548
dirstate-v2: correct documented return values of `pack_dirstate`
Raphaël Gomès <rgomes@octobus.net> [Wed, 26 Oct 2022 12:19:47 +0200] rev 49547
dirstate-v2: fix typos in docstrings
Raphaël Gomès <rgomes@octobus.net> [Fri, 04 Nov 2022 14:52:16 -0400] rev 49546
dirstate-v2: update constant that wasn't kept in sync
Despite the best efforts of the comment, this constant wasn't kept in sync
when the flags were being rewritten.
The fact that this doesn't actually break anything in the Rust implementation
too much (which does use directories) relies on the fact that all nodes
can have children and that dirstate traversal is not based on that flag, but
for metadata in optimizations. However the bug could become more serious should
we start encoding stronger guarantees using a combination of flags including
this one.
Matt Harbison <matt_harbison@yahoo.com> [Tue, 18 Oct 2022 13:56:45 -0400] rev 49545
lfs: avoid closing connections when the worker doesn't fork
Probably not much more than an minor optimization, but could be useful in the
case of `hg verify` where missing blobs are fetched one at a time.
Matt Harbison <matt_harbison@yahoo.com> [Tue, 18 Oct 2022 13:36:33 -0400] rev 49544
lfs: fix blob corruption when tranferring with workers on posix
The problem seems to be that the connection used to request the location of the
blobs is sitting in the connection pool, and then when workers are forked, they
all see and attempt to use the same connection. This garbles everything. I
have no clue how this ever worked reliably (but it seems to, even on Linux, with
SCM Manager 1.58). See previous discussion when worker support was added[1].
It shouldn't be a problem on Windows, since the workers are just threads in the
same process, and can see which connections are marked available and which are
in use. (The fact that `mercurial.keepalive.ConnectionManager.set_ready()`
doesn't acquire a lock does give me some pause though.)
[1] https://phab.mercurial-scm.org/D1568#31621
Matt Harbison <matt_harbison@yahoo.com> [Tue, 18 Oct 2022 12:58:34 -0400] rev 49543
keepalive: add `__repr__()` to the HTTPConnection class to ease debugging
By default, this just printed the class name and memory address. By displaying
the address and ports on both sides of the socket, it makes it easier to figure
out what's in the ConnectionManager, and correlate with WireShark traces.
It looks like the two connections mentioned in the previous commit come about
because the LFS POST request to access the blobs opens connection 1, and gets a
401. Then for some reason, the follow up with credentials opens a new socket,
instead of using the existing one in the pool. I have no clue why.
This can be seen with something like this in the blobstore:
```
for h in self.urlopener.handlers:
if hasattr(h, "close_all"):
print('open connections on %s in pid %d' % (type(h), os.getpid()))
for host, conns in h._cm.get_all().items():
for c in conns:
print('connection: %r' % c)
```
Matt Harbison <matt_harbison@yahoo.com> [Tue, 18 Oct 2022 11:54:58 -0400] rev 49542
keepalive: ensure `close_all()` actually closes all cached connections
While debugging why LFS blob downloads are getting corrupted with workers, I
noticed that prior to spinning up the workers, the ConnectionManager has 2
connections to the server and calling `KeepAliveHandler.close_all()` left one
behind. The reason is the value component of `self._cm.get_all().items()` is a
list, and `self._cm.remove()` modifies said list while the caller is iterating
over it. Now `get_all()` is a deep copy of both the dict and lists in all
cases.
Pierre-Yves David <pierre-yves.david@octobus.net> [Thu, 22 Sep 2022 16:27:17 +0200] rev 49541
tests: remove non-python3 line matching and tests block
We don't support Python2 anymore
Matt Harbison <matt_harbison@yahoo.com> [Wed, 02 Nov 2022 16:46:46 -0400] rev 49540
localrepo: byteify the requirements.DIRSTATE_TRACKED_HINT_Vx warning message
Flagged by PyCharm.
Arseniy Alekseyev <aalekseyev@janestreet.com> [Mon, 31 Oct 2022 16:15:54 +0000] rev 49539
rhg: fallback to slow path on invalid patterns in hgignore
Arseniy Alekseyev <aalekseyev@janestreet.com> [Mon, 31 Oct 2022 16:15:30 +0000] rev 49538
rhg: add a test involving hgignore lookaround
Anton Shestakov <av6@dwimlabs.net> [Mon, 31 Oct 2022 16:50:22 +0400] rev 49537
pywatchman: remove obsolete comments about importing from future
See
6000f5b25c9b.
Anton Shestakov <av6@dwimlabs.net> [Mon, 31 Oct 2022 16:36:00 +0400] rev 49536
demandimport: remove an obsolete comment about importing from future
See
6000f5b25c9b.
Matt Harbison <matt_harbison@yahoo.com> [Wed, 19 Oct 2022 11:10:54 -0400] rev 49535
mr-template: wrap the instructions inside a comment block
At least in preview mode, this hides the text so the user doesn't have to delete
it. It's still visible in edit mode, so the user sees it.
Matt Harbison <matt_harbison@yahoo.com> [Wed, 19 Oct 2022 11:50:40 -0400] rev 49534
revlog: use the user facing filename as the display_id for filelogs
I had trouble isolating some LFS blob corruption detected by `hg verify` because
the traceback referenced a file, but with the `data/` prefix in the `.hg/store`
path, so it couldn't be located with the `file()` revset:
```
Traceback (most recent call last):
File "/mnt/d/mercurial/mercurial/revlog.py", line 3209, in verifyintegrity
_verify_revision(self, skipflags, state, node)
File "/mnt/d/mercurial/hgext/lfs/wrapper.py", line 246, in _verify_revision
orig(rl, skipflags, state, node)
File "/mnt/d/mercurial/mercurial/revlog.py", line 158, in _verify_revision
rl.revision(node)
File "/mnt/d/mercurial/mercurial/revlog.py", line 1816, in revision
return self._revisiondata(nodeorrev, _df)
File "/mnt/d/mercurial/mercurial/revlog.py", line 1870, in _revisiondata
self.checkhash(text, node, rev=rev)
File "/mnt/d/mercurial/mercurial/revlog.py", line 1996, in checkhash
% (self.display_id, pycompat.bytestr(revornode))
mercurial.error.RevlogError: integrity check failed on data/EXE/PPC/shrinksrec.exe:0
```
(I'm a little surprised it resulted in a stacktrace instead of just a message,
but that's a different issue. I'm also not sure how to trigger the simplestore
case, since IIUC, it's also a revlog based store.)
It's not clear how to handle the changelog and manifest (because the user
doesn't interact with them as a file), so those cases are left alone. The other
thing that would be nice to improve somehow is to indicate that the ":0" is a
revlog revision, not the changeset revision that users are used to. I'm not
sure how to handle the "or node" part though.