Mercurial > hg
view tests/test-show-work.t @ 44118:f81c17ec303c
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
author | Gregory Szorc <gregory.szorc@gmail.com> |
---|---|
date | Mon, 20 Jan 2020 23:51:25 -0800 |
parents | 34a46d48d24e |
children |
line wrap: on
line source
$ cat >> $HGRCPATH << EOF > [extensions] > show = > EOF $ hg init repo0 $ cd repo0 Command works on an empty repo $ hg show work Single draft changeset shown $ echo 0 > foo $ hg -q commit -A -m 'commit 0' $ hg show work @ 9f17 commit 0 Even when it isn't the wdir $ hg -q up null $ hg show work o 9f17 commit 0 Single changeset is still there when public because it is a head $ hg phase --public -r 0 $ hg show work o 9f17 commit 0 A draft child will show both it and public parent $ hg -q up 0 $ echo 1 > foo $ hg commit -m 'commit 1' $ hg show work @ 181c commit 1 o 9f17 commit 0 Multiple draft children will be shown $ echo 2 > foo $ hg commit -m 'commit 2' $ hg show work @ 128c commit 2 o 181c commit 1 o 9f17 commit 0 Bumping first draft changeset to public will hide its parent $ hg phase --public -r 1 $ hg show work @ 128c commit 2 o 181c commit 1 ~ Multiple DAG heads will be shown $ hg -q up -r 1 $ echo 3 > foo $ hg commit -m 'commit 3' created new head $ hg show work @ f0ab commit 3 | o 128c commit 2 |/ o 181c commit 1 ~ Even when wdir is something else $ hg -q up null $ hg show work o f0ab commit 3 | o 128c commit 2 |/ o 181c commit 1 ~ Draft child shows public head (multiple heads) $ hg -q up 0 $ echo 4 > foo $ hg commit -m 'commit 4' created new head $ hg show work @ 668c commit 4 | o f0ab commit 3 | | o 128c commit 2 | |/ | o 181c commit 1 |/ o 9f17 commit 0 $ cd .. Branch name appears in output $ hg init branches $ cd branches $ echo 0 > foo $ hg -q commit -A -m 'commit 0' $ echo 1 > foo $ hg commit -m 'commit 1' $ echo 2 > foo $ hg commit -m 'commit 2' $ hg phase --public -r . $ hg -q up -r 1 $ hg branch mybranch marked working directory as branch mybranch (branches are permanent and global, did you want a bookmark?) $ echo 3 > foo $ hg commit -m 'commit 3' $ echo 4 > foo $ hg commit -m 'commit 4' $ hg show work @ f8dd (mybranch) commit 4 o 90cf (mybranch) commit 3 | o 128c commit 2 |/ o 181c commit 1 ~ $ cd .. Bookmark name appears in output $ hg init bookmarks $ cd bookmarks $ echo 0 > foo $ hg -q commit -A -m 'commit 0' $ echo 1 > foo $ hg commit -m 'commit 1' $ echo 2 > foo $ hg commit -m 'commit 2' $ hg phase --public -r . $ hg bookmark @ $ hg -q up -r 1 $ echo 3 > foo $ hg commit -m 'commit 3' created new head $ echo 4 > foo $ hg commit -m 'commit 4' $ hg bookmark mybook $ hg show work @ cac8 (mybook) commit 4 o f0ab commit 3 | o 128c (@) commit 2 |/ o 181c commit 1 ~ $ cd .. Tags are rendered $ hg init tags $ cd tags $ echo 0 > foo $ hg -q commit -A -m 'commit 1' $ echo 1 > foo $ hg commit -m 'commit 2' $ hg tag 0.1 $ hg phase --public -r . $ echo 2 > foo $ hg commit -m 'commit 3' $ hg tag 0.2 $ hg show work @ 3758 Added tag 0.2 for changeset 6379c25b76f1 o 6379 (0.2) commit 3 o a2ad Added tag 0.1 for changeset 6a75536ea0b1 ~ $ cd .. Multiple names on same changeset render properly $ hg init multiplenames $ cd multiplenames $ echo 0 > foo $ hg -q commit -A -m 'commit 1' $ hg phase --public -r . $ hg branch mybranch marked working directory as branch mybranch (branches are permanent and global, did you want a bookmark?) $ hg bookmark mybook $ echo 1 > foo $ hg commit -m 'commit 2' $ hg show work @ 3483 (mybook) (mybranch) commit 2 o 97fc commit 1 Multiple bookmarks on same changeset render properly $ hg book mybook2 $ hg show work @ 3483 (mybook mybook2) (mybranch) commit 2 o 97fc commit 1 $ cd .. Extra namespaces are rendered $ hg init extranamespaces $ cd extranamespaces $ echo 0 > foo $ hg -q commit -A -m 'commit 1' $ hg phase --public -r . $ echo 1 > foo $ hg commit -m 'commit 2' $ echo 2 > foo $ hg commit -m 'commit 3' $ hg --config extensions.revnames=$TESTDIR/revnamesext.py show work @ 32f3 (r2) commit 3 o 6a75 (r1) commit 2 o 97fc (r0) commit 1 Obsolescence information appears in labels. $ cat >> .hg/hgrc << EOF > [experimental] > evolution=createmarkers > EOF $ hg debugobsolete `hg log -r 'desc("commit 2")' -T "{node}"` 1 new obsolescence markers obsoleted 1 changesets 1 new orphan changesets $ hg show work --color=debug @ [log.changeset changeset.draft changeset.unstable instability.orphan|32f3] [log.description|commit 3] x [log.changeset changeset.draft changeset.obsolete|6a75] [log.description|commit 2] ~ $ cd .. Prefix collision on hashes increases shortest node length $ hg init hashcollision $ cd hashcollision $ echo 0 > a $ hg -q commit -Am 0 $ for i in 17 1057 2857 4025; do > hg -q up 0 > echo $i > a > hg -q commit -m $i > echo 0 > a > hg commit -m "$i commit 2" > done $ hg show work @ cfd04 4025 commit 2 o c562d 4025 | o 08048 2857 commit 2 | o c5623 2857 |/ | o 6a6b6 1057 commit 2 | o c5625 1057 |/ | o 96b4e 17 commit 2 | o 11424 17 |/ o b4e73 0 $ cd ..