hgdemandimport: apply lazy module loading to sys.meta_path finders Python's `sys.meta_path` finders are the primary objects whose job it is to find a module at import time. When `import` is called, Python iterates objects in this list and calls `o.find_spec(...)` to find a `ModuleSpec` (or None if the module couldn't be found by that finder). If no meta path finder can find a module, import fails. One of the default meta path finders is `PathFinder`. Its job is to import modules from the filesystem and is probably the most important importer. This finder looks at `sys.path` and `sys.path_hooks` to do its job. The `ModuleSpec` returned by `MetaPathImporter.find_spec()` has a `loader` attribute, which defines the concrete module loader to use. `sys.path_hooks` is a hook point for teaching `PathFinder` to instantiate custom loader types. Previously, we injected a custom `sys.path_hook` that told `PathFinder` to wrap the default loaders with a loader that creates a module object that is lazy. This approach worked. But its main limitation was that it only applied to the `PathFinder` meta path importer. There are other meta path importers that are registered. And in the case of PyOxidizer loading modules from memory, `PathFinder` doesn't come into play since PyOxidizer's own meta path importer was handling all imports. This commit changes our approach to lazy module loading by proxying all meta path importers. Specifically, we overload the `find_spec()` method to swap in a wrapped loader on the `ModuleSpec` before it is returned. The end result of this is all meta path importers should be lazy. As much as I would have loved to utilize .__class__ manipulation to achieve this, some meta path importers are implemented in C/Rust in such a way that they cannot be monkeypatched. This is why we use __getattribute__ to define a proxy. Also, this change could theoretically open us up to regressions in meta path importers whose loader is creating module objects which can't be monkeypatched. But I'm not aware of any of these in the wild. So I think we'll be safe. According to hyperfine, this change yields a decent startup time win of 5-6ms: ``` Benchmark #1: ~/.pyenv/versions/3.6.10/bin/python ./hg version Time (mean ± σ): 86.8 ms ± 0.5 ms [User: 78.0 ms, System: 8.7 ms] Range (min … max): 86.0 ms … 89.1 ms 50 runs Time (mean ± σ): 81.1 ms ± 2.7 ms [User: 74.5 ms, System: 6.5 ms] Range (min … max): 77.8 ms … 90.5 ms 50 runs Benchmark #2: ~/.pyenv/versions/3.7.6/bin/python ./hg version Time (mean ± σ): 78.9 ms ± 0.6 ms [User: 70.2 ms, System: 8.7 ms] Range (min … max): 78.1 ms … 81.2 ms 50 runs Time (mean ± σ): 73.4 ms ± 0.6 ms [User: 65.3 ms, System: 8.0 ms] Range (min … max): 72.4 ms … 75.7 ms 50 runs Benchmark #3: ~/.pyenv/versions/3.8.1/bin/python ./hg version Time (mean ± σ): 78.1 ms ± 0.6 ms [User: 70.2 ms, System: 7.9 ms] Range (min … max): 77.4 ms … 80.9 ms 50 runs Time (mean ± σ): 72.1 ms ± 0.4 ms [User: 64.4 ms, System: 7.6 ms] Range (min … max): 71.4 ms … 74.1 ms 50 runs ``` Differential Revision: https://phab.mercurial-scm.org/D7954

hgdemandimport: disable on Python 3.5 The demand importer functionality isn't working at all on Python 3.5. I'm not sure what's wrong. Since it isn't working, let's disable it completely. ``` $ HGRCPATH= hyperfine -w 1 -r 50 -- "~/.pyenv/versions/3.5.9/bin/python ./hg version" \ "HGDEMANDIMPORT=disable ~/.pyenv/versions/3.5.9/bin/python ./hg version" Benchmark #1: ~/.pyenv/versions/3.5.9/bin/python ./hg version Time (mean ± σ): 163.7 ms ± 2.2 ms [User: 148.5 ms, System: 15.7 ms] Range (min … max): 161.0 ms … 170.2 ms 50 runs Benchmark #2: HGDEMANDIMPORT=disable ~/.pyenv/versions/3.5.9/bin/python ./hg version Time (mean ± σ): 164.3 ms ± 1.4 ms [User: 148.2 ms, System: 16.6 ms] Range (min … max): 161.4 ms … 169.8 ms 50 runs ``` Differential Revision: https://phab.mercurial-scm.org/D7953

py3: suppress unraisable exceptions in test-worker.t Python 3.8 calls sys.unraisablehook when an unraisable exception is encountered. The default behavior is to print a warning. test-worker.t was triggering this hook due to a race between a newly forked process exiting and that process's _os.register_at_fork handlers running. I was seeing the stdlib's random module in the stack re-seeding itself. Although there could be other after-fork handlers in the mix. This commit defines sys.unraisablehook to effectively no-op. This suppresses the warning and makes test output on Python 3.8 consistent with prior versions. test-worker.t now passes on Python 3.8. Differential Revision: https://phab.mercurial-scm.org/D7949

rust: add a README In particular to explain how to build any of the rust. It's neither obvious, nor easy to find out, nor easy to determine if you did it right without some documentation. Differential Revision: https://phab.mercurial-scm.org/D7952

rust: move hgcli's README out of the way My understanding is that it's not meant to be used in the current form. Differential Revision: https://phab.mercurial-scm.org/D7951

verify: avoid spurious integrity warnings in verbose mode (issue6172) The issue seems to revolve around renames in filtered commits, and only occurred in verbose mode. The problem occurs in the `# check renames` stage, around line 577. Without using the unfiltered repo, this test would have printed: $ hg verify -v repository uses revlog format 1 checking changesets checking manifests crosschecking files in changesets and manifests checking files foo@25: checking rename of 71ec0570c325: filtered revision '25' foobar@26: checking rename of 1b549296015b: filtered revision '26' checked 28 changesets with 16 changes to 11 files 2 integrity errors encountered! (first damaged changeset appears to be 25) [1] Differential Revision: https://phab.mercurial-scm.org/D7950

py3: glob over exception in test-check-py3-compat.t Python 3.6+ raise ModuleNotFoundError and older versions raise ImportError. Glob over the exception differences. For whatever reason, we were already doing this for one failure. But not all occurrences of ModuleNotFoundError were changed. Who knows. This test should now pass on all Python versions (although I didn't check Windows). Differential Revision: https://phab.mercurial-scm.org/D7939

py3: string normalization and I/O tweaks in test-lfs.t The print was inserting b'' on Python 3. In addition, since we weren't writing to the ui instance (which isn't readily available in this function), output order could get mixed up. We add some pycompat casts and a stdout flush to make the test happy on all Python versions. Differential Revision: https://phab.mercurial-scm.org/D7938