config: add option to control creation of empty successors during rewrite The default for many history-rewriting commands (e.g. rebase and absorb) is that changesets which would become empty are not created in the target branch. This makes sense if the source branch consists of small fix-up changes. For more advanced workflows that make heavy use of history-editing to create curated patch series, dropping empty changesets is not as important or even undesirable. Some users want to keep the meta-history, e.g. to make finding comments in a code review tool easier or to avoid that divergent bookmarks are created. For that, obsmarkers from the (to-be) empty changeset to the changeset(s) that already made the changes should be added. If a to-be empty changeset is pruned without a successor, adding the obsmarkers is hard because the changeset has to be found within the hidden part of the history. If rebasing in TortoiseHg, it’s easy to miss the fact that the to-be empty changeset was pruned. An empty changeset will function as a reminder that obsmarkers should be added. Martin von Zweigbergk mentioned another advantage. Stripping the successor will de-obsolete the predecessor. If no (empty) successor is created, this won’t be possible. In the future, we may want to consider other behaviors, like e.g. creating the empty successor, but pruning it right away. Therefore this configuration accepts 'skip' and 'keep' instead of being a boolean configuration.

commands: use any() instead of `if a or b or c` Small cleanup for future when we have an option to show configs from shared rc. Differential Revision: https://phab.mercurial-scm.org/D8658

manifest: use the same logic for handling flags in _parse as elsewhere Differential Revision: https://phab.mercurial-scm.org/D8684

manifest: tigher manifest parsing and flag use In the manifest line, flags are put directly after the hash, so the parser has been guessing the presence of flags based on the length of the hash. Replace this assumption by an enumeration of the valid flags and removing them from the hash first as they are distinct input values. Consistently handle the expected 256bit length of the SHA1-replacement in the pure Python parser. Check that setting flags will use one of the blessed values. Extend write logic in the C version to handle 256bit hashes as well. Verify that hashes always have exactly the expected length. Since 1070df141718 we should no longer depend on the old extra-byte hack. Differential Revision: https://phab.mercurial-scm.org/D8679

phases: sparsify phase lists When the internal and archived phase was added, allphase became a large, sparsely populated list. This dramatically increased the number of lookup operations for public relations in `phasecache.phase`. As a first step, define allphases and related lists explicitly to contain only the actual phases. Make phasenames a dictionary and create corresponding dictionaries for mapping phase names back to numbers. Adjust various list to be sparse as well with the exception of phaseroots and phasesets members of phasecache. Keep those as a separate step as it involves changes to the C module. Differential Revision: https://phab.mercurial-scm.org/D8697

phases: updatephases should not skip internal phase Historically, updatephases used to skip the secret phase, but wasn't updated when archived and internal were added. Differential Revision: https://phab.mercurial-scm.org/D8696

repoview: use the phasecache directly to determine mutable revisions This speeds up no-change updates from 1.3s to 0.85s for my NetBSD test repository. Differential Revision: https://phab.mercurial-scm.org/D8670

phases: provide a test and accessor for non-public phase roots This decouples users from the implementation details of the phasecache. Also document a historic artifact about the stored phaseroots. Differential Revision: https://phab.mercurial-scm.org/D8694

hg-core: define a `dirstate_status` `Operation` This is 3/3 in a series of patches to improve dirstate status' code. Following in the footsteps of a46e36b82461, we move the main status functionality to an `Operation`. This will most likely be subject to change in the future (what function signature, what parameters, etc., but we will see when `rhg` gets `hg status` support. Differential Revision: https://phab.mercurial-scm.org/D8663

rust-status: improve documentation and readability This patch is 2/3 in the series to improve the dirstate status code. It adds a number of common type aliases to add more obvious semantics to function signatures, improves/adds documentation where necessary and improves one or two patterns to be more idiomatic. Differential Revision: https://phab.mercurial-scm.org/D8662

rust-status: refactor status into a struct The code for `dirstate/status` has grown too large for comfort, this is the first of three patches that try to improve maintainability. In this patch, refactoring dirstate's status into a struct allows for slimming down function signatures drastically, keeping the mental (and maintenance) burden lower, since pretty much all of them shared a few common arguments. This had the pleasant side-effect of simplifying lifetimes a little. This has no observable impact on performance. The next patch will add/improve documentation and refactor some types. I tried to keep new code down to a minimum in this patch because it's already pretty big. Differential Revision: https://phab.mercurial-scm.org/D8661

tests: avoid "magic" nodeids in test-rebase-legacy.t This helps with readability. Differential Revision: https://phab.mercurial-scm.org/D8735

tests: avoid a "magic" nodeid in test-wireproto-command-lookup.t This helps with readability. Differential Revision: https://phab.mercurial-scm.org/D8734

run-tests: replace '#' with '-' in temp path of repos created for tests If we have multiple cases in a test, that leads us to a temp path of format "<temp-path>-<case1>#<case2>". This leads to hg.parseurl() parsing the path and take part after `#` as a branch name. I encountered this bug while adding support for share-safe case in next patch. Differential Revision: https://phab.mercurial-scm.org/D8647

debugcommands: introduce new debugrequirements command This for now just prints out the list of current requirements. In future this will be helpful in reading requirements from couple of sources, and checking which requirement comes from where. Differential Revision: https://phab.mercurial-scm.org/D8632

scmutil: add writereporequirements() and route requires writing through it In upcoming patches, to implement Share Safe plan we will be introducing requires file in store. We need to route all callers to a single function to check for a share-safe requirement and if present, write requirements to .hg/store/requires instead. After this patch, callers directly calling scmutil.writerequires() are only those where we don't have the repo object, for example when initializing the repository object itself. Differential Revision: https://phab.mercurial-scm.org/D8631

extensions: make `hg nonexistent` not crash with PyOxidizer When running `hg nonexistent`, we try to look for extensions that provide that command. We do that by looking for files in the `hgext.__file__` directory. However, PyOxidizer doesn't provide a `__file__`, so we crash when running with PyOxidizer. We should be able to look for the command in built-in extensions, but we seem to already have code for skipping the scan when running in a frozen binary, so I just modified that code instead. By the way, it also seems like we should be able to search for extensions in the `hgext3rd` module, but we don't do that yet either (before or after this patch). Differential Revision: https://phab.mercurial-scm.org/D8750

procutil: avoid use of deprecated tempfile.mktemp() In the previous version, I used tempfile.mktemp() because it seemed to be the only way to open a file from two processes (the Python documentation says the file backing NamedTemporaryFile can’t be opened a second time on Windows). However, it’s possible when passing the O_TEMPORARY flag to the second open. Source: https://stackoverflow.com/a/15235559/6366251

procutil: make _make_write_all() function private Because this function isn’t meant for general use (e.g., it’s Python 2-only), make in a module-private function by prefixing it with `_`.

dispatch: handle late KeyboardInterrupt occurred in run() User can press Ctrl+C while flushing streams in dispatch.run(). In such case, I think exiting with 255 is better than printing Python traceback and exiting with 1.

dispatch: indent run() function I'll add KeyboardInterrupt handling there.

debuginstall: don't translate encoding messages While Microsoft trained us to read mojibake text, it doesn't make sense to print mojibake messages when debugging the exact issue.

merge with stable

patch: refactor content diffing part in separate fn so extensions can wrap Right now extdiff uses it's own logic using archival to diff two versions of file using external diff tools. This makes the extdiff functionality non-extensible. This series is an attempt to refactor core patch and diff functionality so that extdiff can wrap and reuse it. This will help us in using external diffing tools at more places and not just extdiff command only then. Differential Revision: https://phab.mercurial-scm.org/D8685

tests: check that procutil.std{out,err}.write() returns correct result On Windows, we currently don’t fully test the case when the stream is connected to a TTY, but we test the child process side by connecting them to NUL, which is recognized as a TTY by Python. To make the large write test a bit more useful besides checking that it doesn’t crash, we can check that the write() method returns the correct result.

tests: add tests for when stdout or stderr is connected to `os.devnull` The original motivation was that creating PTYs on Windows is not possible, but `NUL` is recognized as a TTY, so we can have at least some test coverage for the TTY case. I think it doesn’t hurt to run the test cases on all systems.

procutil: ensure that procutil.std{out,err}.write() writes all bytes Python 3 offers different kind of streams and it’s not guaranteed for all of them that calling write() writes all bytes. When Python is started in unbuffered mode, sys.std{out,err}.buffer are instances of io.FileIO, whose write() can write less bytes for platform-specific reasons (e.g. Linux has a 0x7ffff000 bytes maximum and could write less if interrupted by a signal; when writing to Windows consoles, it’s limited to 32767 bytes to avoid the "not enough space" error). This can lead to silent loss of data, both when using sys.std{out,err}.buffer (which may in fact not be a buffered stream) and when using the text streams sys.std{out,err} (I’ve created a CPython bug report for that: https://bugs.python.org/issue41221). Python may fix the problem at some point. For now, we implement our own wrapper for procutil.std{out,err} that calls the raw stream’s write() method until all bytes have been written. We don’t use sys.std{out,err} for larger writes, so I think it’s not worth the effort to patch them.

procutil: move assignments This should probably be part of the previous patch, but folding it results in a less useful word diff, so I decided to keep it separate for review.

procutil: distribute code for stdout It makes sense to have the distinction between Python 2 and 3 at the top level, as we have to fight a different kind of battle on each: On Python 3, we get consistent behavior on all platforms, but need to create correctly-behaving binary streams. On Python 2, we have to account for platform differences.

convert: use bytes for value in extra dict The keys and values of the extra dict are expected to be bytes. Before this fix, there was a crash in mercurial.changelog.encodeextra().