debugcommands: add a `--paranoid` option to `debug-repair-issue-6528` See justification inline. Differential Revision: https://phab.mercurial-scm.org/D11263

repair: improve performance of detection of revisions affected by issue6528 Explanations inside the patch. I've tested this on Mozilla-Central and it's 5 times faster than the naive approach on my laptop. Differential Revision: https://phab.mercurial-scm.org/D11262

debugcommands: introduce a debug command to repair repos affected by issue6528 This command is quite basic and slow, it will loop over the entirety of the filelogs in the repository and check each revision for corruption, then fixes the affected filelogs. It takes under 25 minutes for Mozilla-Central on my not-top-of-the-line laptop, using the `--to-report` and `--from-report` options will make this pretty tolerable to use, I think. This change also introduces a test for the fix. Differential Revision: https://phab.mercurial-scm.org/D11239

contrib: switch the Windows bootstrap environment to py3.9 Use the built in `venv` module instead of `virtualenv` for simplicity, and upgrade to a modern Mercurial that supports py3. One issue here is that `venv` doesn't copy `python3{,Y}.dll` into the `Scripts` subdirectory, so running the `hg.exe` that gets installed immediately fails on a clean system because Python isn't in `PATH`. There is code in `python.exe` to detect when it is in a venv and add the original python install to the DLL lookup path, which we don't do in `hg.exe` yet. The simple workaround for now is to run the `hg` script with `python.exe`. Typically `PYTHONLEGACYWINDOWSSTDIO` must be set in the environment on Windows, but the clone process works without it. Differential Revision: https://phab.mercurial-scm.org/D11275

contrib: comment out the 64-bit py38 dependency installation on Windows Not sure what is going on here, but it appears to not install py3.8 x64 in the usual `C:\hgdev` directory. The x32 installer works fine. I'm assuming this is a bug in this version of the installer, but didn't look into it too much. Differential Revision: https://phab.mercurial-scm.org/D11274

contrib: drop python2.7 from the Windows dependency install script While we aren't quite ready to drop py27 yet, the MS compiler is no longer available from MS (there is a copy on github if it's really needed), and that causes the script to fail. Differential Revision: https://phab.mercurial-scm.org/D11273

contrib: log the command and args for every process installing windows deps This is a little noisier, but makes it simple to debug when things fail. Differential Revision: https://phab.mercurial-scm.org/D11272

windows-ci: run Windows CI automatically alongside the others This will enable us to make Windows Python 3 a first-class citizen for the next 6.0 cycle. We will probably get some flaky tests and we're missing others that are skipped, but we'll turn them on it future patches. Differential Revision: https://phab.mercurial-scm.org/D11256

vfs: always use / as file separator (issue6546) Various part of vfs already enforce `/` usage and using `\` confuse the encoded vfs. So we simply use `/` all the time. Differential Revision: https://phab.mercurial-scm.org/D11260

subrepo: compare normalised vfs path Otherwise the realpath call can turn `/` into `\` on windows confusing the check. (We probably needs this in more location) Differential Revision: https://phab.mercurial-scm.org/D11259

pager: account for flakiness in Windows output This test case is cursed and probably not worth losing more time over. This makes apparent what the intended behavior is while still removing the flakiness from the CI. Differential Revision: https://phab.mercurial-scm.org/D11257

windows-ci: clean up the Heptapod CI file now that the baseline is solid Enough work has been done one the CI side, this now works with little effort on our side. The next patch will remove the manual switch. Differential Revision: https://phab.mercurial-scm.org/D11254

test-nointerrupt: make "sure" the handler "might" trigger (issue6558) We are sure that the signal got sent in the right time frame, however, we still have race, so either the code is actually buggy or we need some security to make sure the signal get processed. We might be affected by https://bugs.python.org/issue43406 ? Differential Revision: https://phab.mercurial-scm.org/D11251

testing: make sure write_file is "atomic" This make sure viewer cannot see the new file with partial content. This was likely the cause of some flakiness in `test-nointerrupt.t` Differential Revision: https://phab.mercurial-scm.org/D11250

test: disable test-subrepo-git.t in python2 + chg I am a couple of days in try to debug that at it seems minor enough with enough other priority to simply disable it for now. Differential Revision: https://phab.mercurial-scm.org/D11253

hgwebdir: avoid systematic full garbage collection Forcing a systematic full garbage collection upon each request can serioulsy harm performance. This is reported as https://bz.mercurial-scm.org/show_bug.cgi?id=6075 With this change we're performing the full collection according to a new setting, `experimental.web.full-garbage-collection-rate`. The default value is 1, which doesn't change the behavior and will allow us to test on real use cases. If the value is 0, no full garbage collection occurs. Regardless of the value of the setting, a partial garbage collection still occurs upon each request (not attempting to collect objects from the oldest generation). This should be enough to take care of reference cycles that have been created by the last request (assessment of this requires changing the setting, not to be 1). In my experience chasing memory leaks in Mercurial servers, the full collection never reclaimed any memory, but this is with Python 3 and biased towards small repositories. On the other hand, as explained in the Python developer docs [1], frequent full collections are very harmful in terms of performance if lots of objects survive the collection, and hence stay in the oldest generation. Note that `gc.collect()` is indeed trying to collect the oldest generation [2]. This happens usually in two cases: - unwanted lingering objects (i.e., an actual memory leak that the GC cannot do anything about). Sadly, we have lots of those these days. - desireable long-term objects, typically in caches (not inner caches carried by repositories, which should be collected with them). This is a subject of interest for the Heptapod project. In short, the flat rate that this change still permits is probably a bad idea in most cases, and the default value can be tweaked later on (or even be set to 0) according to experiments in the wild. The test is inspired from test-hgwebdir-paths.py [1] https://devguide.python.org/garbage_collector/#collecting-the-oldest-generation [2] https://docs.python.org/3/library/gc.html#gc.collect Differential Revision: https://phab.mercurial-scm.org/D11204

check-module-imports: ignore non-stdlib module installed by distribution Previously, the check script would detect breezy as part of the stdlib if installed using the debian package manager. This silence the following complains: tests/test-convert-bzr.t:117: imports not lexically sorted: breezy.bzr.bzrdir < sys tests/test-convert-bzr.t:117: stdlib import "breezy.bzr.bzrdir" follows local import: breezy tests/test-convert-bzr-ghosts.t:7: imports not lexically sorted: breezy.bzr.bzrdir < sys tests/test-convert-bzr-ghosts.t:7: stdlib import "breezy.bzr.bzrdir" follows local import: breezy tests/test-convert-bzr-treeroot.t:7: imports not lexically sorted: breezy.bzr.bzrdir < sys tests/test-convert-bzr-treeroot.t:7: stdlib import "breezy.bzr.bzrdir" follows local import: breezy Differential Revision: https://phab.mercurial-scm.org/D11249

remotefilelog: fix what looks like a wrong refactoring when various store functions started returning a revlog type as the first element of the tuple. Differential Revision: https://phab.mercurial-scm.org/D11243

rust-nodemap: falling back to C impl as mitigation This is a mitigation for https://bz.mercurial-scm.org/show_bug.cgi?id=6554 We see sometimes almost all data except the most recent revisions removed from the persistent nodemap, but we don't know how to reproduce yet. This has sadly repercussions beyond just needing to reconstruct the persistent nodemap: for instance, this automatically filters out all bookmarks pointing to revisions that the nodemap cannot resolve. If such filtering happens in a transaction, the update of the bookmarks file that happens at the end of transaction loses all bookmarks that have been affected. There may be similar consequences for other data. So this is a data loss, something that we have to prevent as soon as possible. As a mitigation measure, we will now fallback to the C implementation in case nodemap lookups failed. This will add some latency, e.g., in discovery, yet less than disabling the persistent nodemap entirely. We considered implementing the fallback directly on the Python side, but `revlog.get_rev()` is not systematically used, there are also several direct calls to the index method (`self.index.rev()` for a `revlog` instance). It is therefore more direct to implement the mitigation in the rust-cpython wrapper. Differential Revision: https://phab.mercurial-scm.org/D11238

obsolete: disable other evolution config options if createmarkers is off We used to raise an abort in this case, but recent changes to local clone command (377d8fc20e34) resulted in destrepo both caring about experimental.evolution config options and not initializing extensions. So imagine if you had evolve and allowdivergence enabled in your ~/.hgrc. Local clone stopped working after 377d8fc20e34 because evolve sets experimental.evolution=all, but only on srcrepo, for destrepo the extension is not initialized. It's possible to make local cloning work by initializing extensions for destrepo in some cases, but in other cases (e.g. allowdivergence in ~/.hgrc, evolve extension in original-repo/.hg/hgrc) it would still fail. In a discussion with Pierre-Yves David it was decided to simply force other evolution options to be false if createmarkers is not enabled. Differential Revision: https://phab.mercurial-scm.org/D11223

fix: use obsolete.isenabled() to check for experimental.allowdivergence Now that obsolete.isenabled() can also check if divergence is allowed, let's use it for consistency. Other experimental.evolution options are already checked via this function. Differential Revision: https://phab.mercurial-scm.org/D11222

rebase: use obsolete.isenabled() to check for experimental.allowdivergence Now that obsolete.isenabled() can also check if divergence is allowed, let's use it for consistency. Other experimental.evolution options are already checked via this function. Differential Revision: https://phab.mercurial-scm.org/D11221

obsolete: disable other evolution config options if createmarkers is off We used to raise an abort in this case, but recent changes to local clone command (377d8fc20e34) resulted in destrepo both caring about experimental.evolution config options and not initializing extensions. So imagine if you had evolve and allowdivergence enabled in your ~/.hgrc. Local clone stopped working after 377d8fc20e34 because evolve sets experimental.evolution=all, but only on srcrepo, for destrepo the extension is not initialized. It's possible to make local cloning work by initializing extensions for destrepo in some cases, but in other cases (e.g. allowdivergence in ~/.hgrc, evolve extension in original-repo/.hg/hgrc) it would still fail. In a discussion with Pierre-Yves David it was decided to simply force other evolution options to be false if createmarkers is not enabled. Differential Revision: https://phab.mercurial-scm.org/D11223

fix: use obsolete.isenabled() to check for experimental.allowdivergence Now that obsolete.isenabled() can also check if divergence is allowed, let's use it for consistency. Other experimental.evolution options are already checked via this function. Differential Revision: https://phab.mercurial-scm.org/D11222

rebase: use obsolete.isenabled() to check for experimental.allowdivergence Now that obsolete.isenabled() can also check if divergence is allowed, let's use it for consistency. Other experimental.evolution options are already checked via this function. Differential Revision: https://phab.mercurial-scm.org/D11221

heptapod-ci: enable pytype testing for all pipelines Now that pytype agrees with the codebase, let's run it by default to catch regressions. Differential Revision: https://phab.mercurial-scm.org/D11242

pytype: add assertions to explain revlogv2 invariants to Pytype Differential Revision: https://phab.mercurial-scm.org/D11241

typing: add several assertions to dirstatemap to appease pytype (grafted from default to stable) I think it's been mentioned in IRC that these can't be None in this case. This fixes: File "/mnt/c/Users/Matt/hg/mercurial/dirstatemap.py", line 213, in addfile: unsupported operand type(s) for &: 'None' and 'int' [unsupported-operands] No attribute '__and__' on None or '__rand__' on int Called from (traceback): line 290, in reset_state File "/mnt/c/Users/Matt/hg/mercurial/dirstatemap.py", line 214, in addfile: unsupported operand type(s) for &: 'None' and 'int' [unsupported-operands] No attribute '__and__' on None or '__rand__' on int Called from (traceback): line 290, in reset_state Differential Revision: https://phab.mercurial-scm.org/D11235

windows: avoid a bytes vs unicode crash reading passwords on py2 This broke in 5b3513177f2b. Specifically, after typing in the password on py2, it would crash with: TypeError: putwch() argument 1 must be cannot convert raw buffers, not str

streamclone: ensure the server sends the right amount of data Otherwise, the client would fail with some confusing error. I have seen an error which I think is this, perhaps due to a concurrent revlog split, which streamclones do not handle correctly and would result in a short read of the index of the revlog being split. Differential Revision: https://phab.mercurial-scm.org/D11236