dirstate: make it mandatory to provide parentfiledata in `set_clean` Gathering the mode, size and mtime, independently from determining that the file is clean is a race-machine. So we just make these information required arguments. (note that the data is still gathered in a racy way in practice, but at least the API is no longer encouraging it.) Differential Revision: https://phab.mercurial-scm.org/D11789

dirstate: do no use `set_clean` in revert The current `set_clean` usage is racy (the file might be modified between its restoration and the `set_clean` call). So we simply leave the file as ambiguous and the next status will fix that. We still have to make sure the copy information is dropped, so we teach dirstate how to do that. The win32txt extension is confused after this because current logic is broken in more location. However this series will ultimately fix that so we "ignore" it for now. Fixing it now is complicated without some extra fix landing later. Differential Revision: https://phab.mercurial-scm.org/D11788

status: adapt the "keyword" extensions to gather stats at lookup time See main core code for details. We don't factor the code in a common function yet, because we will have to adapt a bit more things in the keyword case at the end of the series. Differential Revision: https://phab.mercurial-scm.org/D11787

status: adapt largefile to gather stats at lookup time See the core code for details of why we are doing this. We don't factor the code in a common function yet, because we will have to adapt a bit more things in the largefile case at the end of the series. Differential Revision: https://phab.mercurial-scm.org/D11786

status: gather fixup info at comparison time This is still racy, but on a much small windows. In addition, the API now make it possible for it to not be racy. This also unlock other cleanups that we are about to do regarding mtime ambiguity at gathering time. Differential Revision: https://phab.mercurial-scm.org/D11785

update: filter the ambiguous mtime in update directly Right now, this filtering is done by `dirstate.write` using the time of `dirstate.write` method call. However that filtering is done "too late" It works "fine" as most command are "fast enough", and race rare enough. We are about to change the mtime filtering logic in the dirstate to be more accurate and reliable. However `hg update` will still need such filtering (mostly because it is actually quite racy, even with the existing filtering). So we explicitly implement a similar logic here. Before removing the older one later in the series. Differential Revision: https://phab.mercurial-scm.org/D11784

dirstate: move "get fs now" in the timestamp utility module We will need it during update. Differential Revision: https://phab.mercurial-scm.org/D11783

dirstate-item: allow mtime to be None in "parentdata" This will be useful to filter out unreliable mtime. Differential Revision: https://phab.mercurial-scm.org/D11782

dirstate: add a comment about a racy piece of code during updates This is a bit that is not really correct but works "fine" in practice. Let us write the details down so that people stop wondering how that logic might be correct… It is not. Differential Revision: https://phab.mercurial-scm.org/D11781

tests: add missing `head` for when things go wrong See comment above the changed line, you can get a millions of line of output in case of failure. When don't need to have them all. Differential Revision: https://phab.mercurial-scm.org/D11780

tests: ensure a status will have non ambiguous mtime in some race test For the test to work, we need some mtime to be recorded. For them to be recorded, they need to be "clearly in the past", otherwise edit with the same mtime would be possible. Strictly speaking there might be file system with a minimal mtime increment longer than one second. However it is unlikely that we will run the test on them for now. We can be smarter about this in the future if it becomes necessary. Differential Revision: https://phab.mercurial-scm.org/D11779

tests: make sure no ambiguities remains after the commit This will help to stabilize part of the test that are not relevant for what is actually tested. Differential Revision: https://phab.mercurial-scm.org/D11778

dirstate: clarify a `hg update` invocation in a test It is common for readers of that test to confuse the `hg co` call with a `hg commit`, while it actually means `hg checkout`, an alias for the more common `hg update. So let us use the clearer version. Differential Revision: https://phab.mercurial-scm.org/D11777

test: mark rhg output as flaky rhg is not updating the dirstate on status yet, which make this part of the test flaky. This will be fixed soon. Differential Revision: https://phab.mercurial-scm.org/D11833

chistedit: explain which order the commits are presented in It's not obvious which order the commits in chistedit (and text-based histedit), so let's add a note about it. Differential Revision: https://phab.mercurial-scm.org/D11832

errors: use detailed error for invalid commit-extras argument Differential Revision: https://phab.mercurial-scm.org/D11831

errors: use detailed exit code in pathauditor Differential Revision: https://phab.mercurial-scm.org/D11830

errors: use detailed exit code for RepoLookupError Differential Revision: https://phab.mercurial-scm.org/D11829

errors: use detailed exit code for detected case-collision Differential Revision: https://phab.mercurial-scm.org/D11828

errors: use detailed exit code when trying to merge file outside narrowspec Differential Revision: https://phab.mercurial-scm.org/D11827

errors: use detailed exit code for non-integer number of diff context lines Differential Revision: https://phab.mercurial-scm.org/D11826

patch: add hint about mangled whitespace on bad patch One of the most common reasons that a patch doesn't apply is because its whitespace has been mangled (e.g. by their mail client or though copy&paste). Let's provide a hint about that. Differential Revision: https://phab.mercurial-scm.org/D11825

errors: return more detailed errors when failing to parse or apply patch This patch adds subclasses of `PatchError` so we can distinguish between failure to parse a patch from failure to apply it. It updates the callers to raise either `InputError` or `StateError` depending on which type of error occurred. Differential Revision: https://phab.mercurial-scm.org/D11824

extensions: add a default "*" suboptions prefix This is similar to what we do in other section (e.g. `paths`) and allow to change the behavior for all extensions. Sub options on individual extensions overwrite the default settings. Differential Revision: https://phab.mercurial-scm.org/D11823

extension: add a `required` suboption to enforce the use of an extensions If `required` is set, failing to load an extensions will abort. See the test and documentation for details. Differential Revision: https://phab.mercurial-scm.org/D11822

extensions: highlight the name of the faulty extensions in the error message This make it easier to understand the message when the extensions name is common. Differential Revision: https://phab.mercurial-scm.org/D11821

extensions: refactor handling of loading error make it reusable We will need this in the next patch. Differential Revision: https://phab.mercurial-scm.org/D11820

extensions: ignore "sub-options" when looking for extensions config suboptions are separated by ":" (see the path one for example). So we dont want to confuse these config with actual extensions. We don't have extensions sub option yet, but I am about to introduce one for making sure an extensions can load. So lets level the floor first. Differential Revision: https://phab.mercurial-scm.org/D11819

automation: support Python 3.10 on Windows Python 3.10 is out and we should support it. This commit teaches the automation code to install and support building for Python 3.10. Differential Revision: https://phab.mercurial-scm.org/D11776

contrib: update Windows environment to Python 3.9.9 Let's keep the environment modern. Differential Revision: https://phab.mercurial-scm.org/D11775

rhg: implement the debugignorerhg subcommand This can be used to inspect the generated pattern, but also to benchmark the time it takes to parse hgignore. Differential Revision: https://phab.mercurial-scm.org/D11722

rhg: refactor to use IgnoreFnType alias more widely Differential Revision: https://phab.mercurial-scm.org/D11818

rhg: only complain about poorly configured fallback when falling back Differential Revision: https://phab.mercurial-scm.org/D11751

packaging: bump pygit2 to 1.7.1 Differential Revision: https://phab.mercurial-scm.org/D11805

packaging: bump windows_curses to 2.3.0 This is required for Python 3.10 support. Differential Revision: https://phab.mercurial-scm.org/D11804

packaging: regenerate the requirements files with pip-tools 6.4.0 Somewhere along the line, the formatting changed. There's no change in package content here- it's just some minor text changes. Py2 doesn't have any packages we'll be updating, so I'm not bothering there. Differential Revision: https://phab.mercurial-scm.org/D11803

rhg: Add support for `rhg status -n` The `RHG_STATUS=1` bit added here can be removed when `unset RHG_STATUS` near the top of the file is removed. Differential Revision: https://phab.mercurial-scm.org/D11815

filemerge: simplify slightly by using filectx.decodeddata() Differential Revision: https://phab.mercurial-scm.org/D11802

filemerge: rename _formatconflictmarker() since it formats a label `_formatconflictmarker()`'s name made me think that it would create something like `<<<<<<<` and maybe some more stuff after it, but it's actually just the label that comes after on the same line. So let's rename it to `_formatlabel()`. That's better aligned with the variables the result is assigned it. Differential Revision: https://phab.mercurial-scm.org/D11801

tests: fix test-fix on NetBSD Differential Revision: https://phab.mercurial-scm.org/D11816

rhg: Fix status desambiguation of symlinks and executable files Differential Revision: https://phab.mercurial-scm.org/D11774

rhg: Rename cat_file_is_modified It hasn’t been based on the "cat operation" for some time already. Differential Revision: https://phab.mercurial-scm.org/D11773

rhg: Also parse flags in the manifest parser Differential Revision: https://phab.mercurial-scm.org/D11772

rhg: Propogate manifest parse errors instead of panicking The Rust parser for the manifest file format is an iterator. Now every item from that iterator is a `Result`, which makes error handling required in multiple new places. This makes better recovery on errors possible, compare to a run time panic. Differential Revision: https://phab.mercurial-scm.org/D11771

tests: Adapt test-basic.t expected output for rhg Differential Revision: https://phab.mercurial-scm.org/D11769

rhg: Colored output is not supported Fallback if it is requested explicitly. The default is documented as use color "whenever it seems possible". rhg proceeds without color in that case. Differential Revision: https://phab.mercurial-scm.org/D11762

rhg: [encode] and [decode] config sections are not supported Differential Revision: https://phab.mercurial-scm.org/D11761

rhg: Config commands.status.terse is not supported Differential Revision: https://phab.mercurial-scm.org/D11760

rhg: Propagate config errors in `rhg status` This code was calling `Result::unwrap_or` instead of `Option::unwrap_or` as was presumably intended. Differential Revision: https://phab.mercurial-scm.org/D11759

rhg: $HG_PENDING is not supported Trigger fallback in that case, if configured to do so. Differential Revision: https://phab.mercurial-scm.org/D11758

rhg: Enable `rhg status` in most tests This subcommand is disabled by default because of bugs that make some test fail. Enable it in the rest of the tests in order to avoid regressing them. As with `RHG_ON_UNSUPPORTED`, an environment variable is used instead of a configuration file and `HGRCPATH` because some tests override `HGRCPATH`. Running `unset RHG_STATUS` at the start of a test restores the default of `rhg status` being disabled. Hopefully it can be increasingly removed from test files as bugs are fixed. Differential Revision: https://phab.mercurial-scm.org/D11756

Added signature for changeset 5d08b289e2e5

Added tag 6.0 for changeset 5d08b289e2e5

packaging: make rpm relocatable This makes it possible to install multiple Mercurial rpms on the same machine in different locations. Differential Revision: https://phab.mercurial-scm.org/D11768

dockerrpm: always use current user and group Using uid 1000 and gid 1000 works by default in some cases, but 'id -u' and 'id -g' should work in all cases. Differential Revision: https://phab.mercurial-scm.org/D11767

cext: fix Python 3.11 compatibility - Py_SIZE is not an lvalue (issue6610) Py_SIZE was made a static inline function during Python 3.10 development, as described on https://vstinner.github.io/c-api-opaque-structures.html . e92ca942ddca updated the Mercurial code base accordingly, but somehow missed a couple of cases introduced long time ago in a8c948ee3668. The Python change was dropped for 3.10, but is coming back again in 3.11 .

branching: merge stable into default

heptapod-ci: re-introduce Windows CI We have Windows runners for now, and we're trying to find more, so let's share the joy of flaky tests with everyone. Differential Revision: https://phab.mercurial-scm.org/D11765

cext: define S_IFLNK on Python 2.7 and Windows Before this change, building on Python 2.7 on Windows fails due to S_IFLNK being undefined. This regression was introduced by a32a96079e2d / D11518. It worked on Python 3 because its pyport.h (which is included via Python.h) contains effectively the same code as added by this changeset. Differential Revision: https://phab.mercurial-scm.org/D11763

tests: dd status=noxfer is not portable (issue6523) Differential Revision: https://phab.mercurial-scm.org/D11754

tests: use more portable mkdir -p instead of --parents (issue6591) Differential Revision: https://phab.mercurial-scm.org/D11755

tests: fix test-censor by improving regexp (issue6585) Differential Revision: https://phab.mercurial-scm.org/D11753

ci: stop running with Python 2 We're dropping support for Python 2 after hg 6.0. Differential Revision: https://phab.mercurial-scm.org/D11748

internals: typo pass on the dirstate-v2 help file Differential Revision: https://phab.mercurial-scm.org/D11747

dirstate-v2: fix confusion between requirement and format config variable None of those outputs failed on the CI because dirstate-v2 is not fully tested yet. This also updates the documentation. Differential Revision: https://phab.mercurial-scm.org/D11746

internals: correct dirstate-v2 format variable in the documentation Differential Revision: https://phab.mercurial-scm.org/D11745

hghave: update dirstate-v2 format config option It wasn't kept in sync with the new change in format name. Differential Revision: https://phab.mercurial-scm.org/D11744

merge: with stable

Added signature for changeset a44bb185f6bd

Added tag 6.0rc0 for changeset a44bb185f6bd

merge with default

tests: accept alternative privileged port allocation failure This registers an additional failure message on failed privileged port allocation, equally funcionally valid but previously not handled and causing the test to fail when run in the NixOS sandbox. Differential Revision: https://phab.mercurial-scm.org/D11741

tests: allow the true command to be provided by coreutils The `true` command is sometimes provided as a symbolic link to the `coreutils` single binary. This is the case on NixOS, on which the test was failing because the symbolic link fully resolves to the latter name, equally valid but not previously accepted by the test's golden output. Differential Revision: https://phab.mercurial-scm.org/D11740

tests: add guard check for suid support This adds a check for setuid and setgid support and guards a part of a test requiring this particular filesystem feature. Such feature is notably missing in the Nix/NixOS build sandbox and was causing the whole test to fail for the mercurial package. Differential Revision: https://phab.mercurial-scm.org/D11739

rhg: more efficient `HgPath::join` This commit makes `HgPath::join` slightly more efficient by avoiding one copy. It also avoids a particularly inefficient (quadratic) use of `HgPath::join` by using a new mutating function `HgPathBuf::push` instead. The name for `HgPathBuf::push` is chosen by analogy to `PathBuf::push`. Differential Revision: https://phab.mercurial-scm.org/D11721

relnotes: update 6.0 with the last changes for the release Differential Revision: https://phab.mercurial-scm.org/D11742

rhg: Fix `rhg status` file content comparison This is only used when a file’s metadata make its status ambiguous, which depends on timing of previous command executions. Differential Revision: https://phab.mercurial-scm.org/D11743

rhg: make it possible to opt out of `rhg cat` The reason an opt-out is needed is that there are still behavior differences between `rhg cat` and `hg cat`: - it does not interpret relative paths correctly - it does not interpret patterns correctly, e.g. 're:foobar$' would be interpreted as a verbatim filename - it does not implement the correct semantics of relpath matcher: if given a directory, `hg` concatenates all files in this directory, while `rhg` simply complains Differential Revision: https://phab.mercurial-scm.org/D11723

tests: run the whole hg-core/path_auditor test in a clean temp dir This makes the whole test happen in a clean temporary directory in `$TMPDIR/$random`, and simplifies the test a bit by eliminating unnecessarily dynamic path elements computations. Before this patch, the first part of the test was happening in `/tmp` itself. This allowed coincidentally named files placed in that directory to impact the outcome of the test. Additionally, this made the second part of the test fail on systems on which `$TMPDIR != /tmp`, because the inspected directory was different from the one in which the mock files were being written. This fully fixes the issue only partially solved in db2bc9e667a1. Differential Revision: https://phab.mercurial-scm.org/D11738

fsmonitor: fix traceback after removal of nonnormalset (issue6600) This basically does the same thing it used to do, but in the only place it's needed. Differential Revision: https://phab.mercurial-scm.org/D11707

fsmonitor: fix traceback about bytes and str incompatibility This didn't work anymore in Python 3, switching to a `repr` based reporting does Differential Revision: https://phab.mercurial-scm.org/D11708

backout: backed out changeset 67d14d4e036c Same as the previous patch, this breaks the Windows CI for some yet unknown reason. Differential Revision: https://phab.mercurial-scm.org/D11727

backout: backed out changeset f78d8b8c46d7 This and the following backout exist because the original patches break the Windows CI for some yet unknown reason. Differential Revision: https://phab.mercurial-scm.org/D11726

relnotes: clear next release notes Differential Revision: https://phab.mercurial-scm.org/D11737

relnotes: add 6.0 release notes Differential Revision: https://phab.mercurial-scm.org/D11736

relnotes: fix wrong command name in 5.9 relnotes Differential Revision: https://phab.mercurial-scm.org/D11735

pyoxidizer: use in-memory resources on non-Windows platforms In-memory resources were disabled for macOS in 7bc1beed, and for all platforms in c900d962. Unfortunately this made it so that we were no longer producing standalone binaries on these platforms, and would have to ship the .py and .pyc files alongside the pyoxidized binary. These changes are no longer necessary after f6b04591, which disabled pep517 and solved the issue we were encountering. Differential Revision: https://phab.mercurial-scm.org/D11734

compat: don't rely on cpython-specific builtins manipulation Pierre Augier signaled on the mailing list that this fails on Pypy and pointed out the correct solution. Differential Revision: https://phab.mercurial-scm.org/D11732

rust-format: update rustfmt version This hasn't been updated in a while, and since we're updating the MSRV to 1.48.0, we might as well do this. Differential Revision: https://phab.mercurial-scm.org/D11733

rust: update the minimum version of Rust Debian Bullseye has just been released, and it carries `rustc 1.48.0`. This actually implies a regression that we can't really do anything about in `rhg`. See https://github.com/rust-lang/rust/issues/88825. In short, closed (or bad) standard file descriptors are reopened silently with no way of telling by the Rust runtime before `main` is executed. This means that closed fds are not forwarded to the subprocess we run in case of fallback. This is a bit sad, but probably not something worth worrying too much about. Differential Revision: https://phab.mercurial-scm.org/D11341

dirstate-v2: freeze the on-disk format It seems the format as reached a good balance. With a core of new capabilities that motivated it initially and enough new feature and room for future improvement to be a clear progress we can set a milestone for. Having the format frozen will help the feature to get real life testing, outside of the test suite. The feature itself stay experimental but the config gains a new name to avoid people enable non-frozen version by default. If too many bugs are reported during the RC we might move the format back to experimental and drop its support in future version (in favor of a new one) Differential Revision: https://phab.mercurial-scm.org/D11709

merge: with stable

windows: use raw string in test log paths Otherwise the backslashes will work as escapes on Windows and break everything. Differential Revision: https://phab.mercurial-scm.org/D11725

pull: make the new argument a keyword argument As per feedback from Dan Villiom Podlaski Christiansen in https://phab.mercurial-scm.org/D11674#179866 Differential Revision: https://phab.mercurial-scm.org/D11730

dirstate: make sure that status does not overlook the fallback flags Without this extra checks, file with fallback flags change as the only change would be overlooked. In the future we might store proper data in the dirstate and do less lookup. However, for now this will do to make sure that 6.0 is forward compatible with later version. Differential Revision: https://phab.mercurial-scm.org/D11729

dirstate: use a single closure for get_flags The previous code was overlooking fallback when neither symlink not exec was supported. The number of "variants" is getting too high, so I am consolidating this in a single closure that should be easier to maintains. This also ensure that fallback flags are always taken into account. (they are not user code yet, but small experimentation shown that the feature was working as intended.) A a small side effect we need to check for symlink support more lazily and this show up in the test in a couple of places. Differential Revision: https://phab.mercurial-scm.org/D11728