record: when backing up, avoid generating very long filenames If the original file's path is longer than the individual filename maximum length (256 on Linux, I believe?), then this mechanism of "replace slashes with underscores" causes an error. Now, we'll produce just the "basename" of the file, plus some stuff to ensure it's unique. This can be potentially confusing for users if there's a file with the same name in multiple directories, but I suspect that this is better than just breaking. Example: `<reporoot>/a/long/path/to/somefile.txt` used to be backed up as `<reporoot>/.hg/record-backups/a_long_path_to_somefile.txt.abcdefgh`, it will now be backed up as `<reporoot>/.hg/record-backups/somefile.txt.abcdefgh` We could do the naive thing (what we were doing before) and have it to doing something with either subdirectories (`<backuproot>/a/long/path/to/somefile.txt.abcdefgh` or minimize #dirs with `<backuproot>/a_long_path/to_somefile.txt.abcdefgh`), prefix-truncated paths (such as `<backuproot>/__ath_to_somefile.txt.abcdefgh`, where that `__` elides enough to get us under 255 chars (counting the +9 we need to add!)), or hash-of-dirname (`<backuproot>/<sha1sum_of_dirname>/somefile.txt.abcdefgh`), but ultimately every option felt over engineered and that it would be more likely to cause problems than it would be to solve any, especially if it was conditional on directory length. Differential Revision: https://phab.mercurial-scm.org/D9207

mergestate: add `allextras()` to get all extras `extras()` can only be used for getting extra for a file. However at couple of places in code, we wanted to iterate over all the extras stored with the mergestate and they were accessing the private `_stateextras`. We add a new function for this. Differential Revision: https://phab.mercurial-scm.org/D9190

mergestate: document `o` merge record state in _mergestate_base docs _mergestate_base documentation serves as a nice documentation for mergestate. This also documents known merge records and known merge record states. I missed adding `o` state to it when I introduced it. Let's add it now. Differential Revision: https://phab.mercurial-scm.org/D9156

fix: update commit hash references in the new commits Differential Revision: https://phab.mercurial-scm.org/D9183

absorb: update commit hash references in the new commits Differential Revision: https://phab.mercurial-scm.org/D9182

rewriteutil: handle dropped commits when updating description hashes In looking to leverage this with the absorb extension, the old -> new mapping there allows the new value to be None. We could filter that out and not pass it to this method, but it seems worth a message to the user. (I wonder if these should be an info or warning, because it's unlikely people are using `-v` regularly.) Differential Revision: https://phab.mercurial-scm.org/D9181

dirstate-tree: move a conditional in an explicit boolean This will help readability a bit and make the next change simpler to read. Differential Revision: https://phab.mercurial-scm.org/D9202

rust: cleanup some white space in a dock They seems to have sneaked there somehow. Differential Revision: https://phab.mercurial-scm.org/D9201

commit: pass ChangingFiles object as argument to _process_files Instead of returning it, we pass it as an argument. This makes the whole if-else in `_prepare_files` a bit simpler. Else each if-else branch was creating the object. Differential Revision: https://phab.mercurial-scm.org/D9194

commit: pass mergestate into `_process_files` instead of re-reading it Differential Revision: https://phab.mercurial-scm.org/D9193

commit: move salvaged calculation a bit earlier in the function This helps us initialize mergestate before and now we can pass it into `_process_files()` instead of re-reading it there. Differential Revision: https://phab.mercurial-scm.org/D9192

commit: refactor salvage calculation to a different function Differential Revision: https://phab.mercurial-scm.org/D9191

tests: run test-copies-chain-merge.t also with copies in changesets We have these tests already and it seems like a waste to not run them in the changesets case. The biggest differences stem from `hg log --follow` not working with copies stored in the changeset extras. Differential Revision: https://phab.mercurial-scm.org/D9173

tests: add test of copies suggested by Pierre-Yves on D9159 Differential Revision: https://phab.mercurial-scm.org/D9171

tests: update test-copies-chain-merge.t to not use empty files Merging empty files is not very interesting or realistic. Differential Revision: https://phab.mercurial-scm.org/D9172

revset: fix sorting key of wdir revision It would go wrong on Python 2, and would crash on Python 3.

grep: move prep() to grepsearcher class

grep: move readfile() to grepsearcher class

grep: move getbody() to grepsearcher class

grep: add stub class that maintains cache and states of grep operation Prepares for extracting stateful functions from commands.grep().

grep: move match and diff logic to new module commands.grep() has lots of functions and classes. Let's split it into reusable components so we can leverage them to implement a revset predicate for 'hg grep --diff'. I want to do 'hg log -r "diff(pattern)"'.

grep: explicitly pass regexp to closure functions These functions will be extracted to new module.

upgrade: improve documentation of matchrevlog() It was not clear what selected for cloning meant. I updated with an extra line of description.

automation: support running against Python 3.9 We just added CPython 3.9.0 to the automation environment. Let's support actions targeting it. One side-effect of this change is that we'll produce Windows wheels for Python 3.9 by default. This is desirable. Differential Revision: https://phab.mercurial-scm.org/D9189

automation: upgrade packages in Linux environment It's been a while since we've ran `pip-compile -U`. Let's do that. When I ran this command, black was upgraded and newer versions of black complains about formatting differences. I didn't feel like shaving a yak, so I pinned the current black version to avoid churn. I manually ran tests against all major Python versions and I'm fairly certain no new test failures were introduced. Differential Revision: https://phab.mercurial-scm.org/D9188

packaging: upgrade packages in Windows environment I ran `pip-compile -U` to update all packages to latest versions. I did this so the Windows environment is more modern. dulwich 0.20 dropped support for Python 2.7. Rather than add yet another requirements.txt file, I decided to just pin at an older release to minimize complexity. Differential Revision: https://phab.mercurial-scm.org/D9187

contrib: install Python 3.9.0 We update the Windows and Linux environments to install CPython 3.9.0, which was released a few days ago. Differential Revision: https://phab.mercurial-scm.org/D9186

contrib: stop installing Python 3.5 and 3.6 in Windows environment We're only publishing Python 3.7+ wheels and other builds on Windows. Python 3.5 and 3.6 are not supported on this platform. Let's stop installing them in the development environment. Differential Revision: https://phab.mercurial-scm.org/D9185

py3: convert an exception message to bytes One of our users ran into this (unrelated to my recent work on more specific errors). Differential Revision: https://phab.mercurial-scm.org/D9184

tests: fix test-url.py on py3, broken by D9179 Differential Revision: https://phab.mercurial-scm.org/D9180

revset: add a `node` key for sorting Sorting by node give an arbitrary, but stable order which is useful for sampling.

pyoxidizer: produce working Python 3 Windows installers (issue6366) While we've had code to produce Python 3 Windows installers with PyOxidizer, we haven't been advertising them on the web site due to a bug in making TLS connections and issues around resource handling. This commit upgrades our PyOxidizer install and configuration to use a recent Git commit of PyOxidizer. This new version of PyOxidizer contains a *ton* of changes, improvements, and bug fixes. Notably, Windows shared distributions now mostly "just work" and the TLS bug and random problems with Python extension modules in the standard library go away. And Python has been upgraded from 3.7 to 3.8.6. The price we pay for this upgrade is a ton of backwards incompatible changes to Starlark. I applied this commit (the overall series actually) on stable to produce Windows installers for Mercurial 5.5.2, which I published shortly before submitting this commit for review. In order to get the stable branch working, I decided to take a less aggressive approach to Python resource management. Previously, we were attempting to load all Python modules from memory and were performing some hacks to copy Mercurial's non-module resources into additional directories in Starlark. This commit implements a resource callback function in Starlark (a new feature since PyOxidizer 0.7) to dynamically assign standard library resources to in-memory loading and all other resources to filesystem loading. This means that Mercurial's files and all the other packages we ship in the Windows installers (e.g. certifi and pygments) are loaded from the filesystem instead of from memory. This avoids issues due to lack of __file__ and enables us to ship a working Python 3 installer on Windows. The end state of the install layout after this patch is not ideal for @: we still copy resource files like templates and help text to directories next to the hg.exe executable. There is code in @ to use importlib.resources to load these files and we could likely remove these copies once this lands on @. But for now, the install layout mimics what we've shipped for seemingly forever and is backwards compatible. It allows us to achieve the milestone of working Python 3 Windows installers and gets us a giant step closer to deleting Python 2. Differential Revision: https://phab.mercurial-scm.org/D9148