perf: clear vfs audit_cache before each run
When generating a stream clone, we spend a large amount of time auditing path.
Before this changes, the first run was warming the vfs cache for the other
runs, leading to a large runtime difference and a "faulty" reported timing for
the operation.
We now clear this important cache between run to get a more realistic timing.
Below are some example of median time change when clearing these cases. The
maximum time for a run did not changed significantly.
### data-env-vars.name = mozilla-central-2018-08-01-zstd-sparse-revlog
# benchmark.name = hg.perf.exchange.stream.generate
# bin-env-vars.hg.flavor = default
# bin-env-vars.hg.py-re2-module = default
# benchmark.variants.version = latest
no-clearing: 17.289905
cache-clearing: 21.587965 (+24.86%, +4.30)
## data-env-vars.name = mozilla-central-2024-03-22-zstd-sparse-revlog
no-clearing: 32.670748
cache-clearing: 40.467095 (+23.86%, +7.80)
## data-env-vars.name = mozilla-try-2019-02-18-zstd-sparse-revlog
no-clearing: 37.838858
cache-clearing: 46.072749 (+21.76%, +8.23)
## data-env-vars.name = mozilla-unified-2024-03-22-zstd-sparse-revlog
no-clearing: 32.969395
cache-clearing: 39.646209 (+20.25%, +6.68)
In addition, this significantly reduce the timing difference between the
performance command, from the perf extensions and a `real `hg bundle` call
producing a stream bundle. Some significant differences remain especially on
the "mozilla-try" repositories, but they are now smaller.
Note that some of that difference will actually not be
attributable to the stream generation (like maybe phases or branch map
computation).
Below are some benchmarks done on a currently draft changeset fixing some
unrelated slowness in `hg bundle` (34a78972af409d1ff37c29e60f6ca811ad1a457d)
### data-env-vars.name = mozilla-central-2018-08-01-zstd-sparse-revlog
# bin-env-vars.hg.flavor = default
# bin-env-vars.hg.py-re2-module = default
hg.perf.exchange.stream.generate: 21.587965
hg.command.bundle: 24.301799 (+12.57%, +2.71)
## data-env-vars.name = mozilla-central-2024-03-22-zstd-sparse-revlog
hg.perf.exchange.stream.generate: 40.467095
hg.command.bundle: 44.831317 (+10.78%, +4.36)
## data-env-vars.name = mozilla-unified-2024-03-22-zstd-sparse-revlog
hg.perf.exchange.stream.generate: 39.646209
hg.command.bundle: 45.395258 (+14.50%, +5.75)
## data-env-vars.name = mozilla-try-2019-02-18-zstd-sparse-revlog
hg.perf.exchange.stream.generate: 46.072749
hg.command.bundle: 55.882608 (+21.29%, +9.81)
## data-env-vars.name = mozilla-try-2023-03-22-zlib-general-delta
hg.perf.exchange.stream.generate: 334.716708
hg.command.bundle: 377.856767 (+12.89%, +43.14)
## data-env-vars.name = mozilla-try-2023-03-22-zstd-sparse-revlog
hg.perf.exchange.stream.generate: 302.972301
hg.command.bundle: 326.098755 (+7.63%, +23.13)
== New Features ==
* `hg purge`/`hg clean` can now delete ignored files instead of
untracked files, with the new -i flag.
* `hg pull` now has a `--confirm` flag to prompt before applying changes.
Config option `pull.confirm` is also added for that.
* `hg log` now defaults to using an '%' symbol for commits involved
in unresolved merge conflicts. That includes unresolved conflicts
caused by e.g. `hg update --merge` and `hg graft`. '@' still takes
precedence, so what used to be marked '@' still is.
* New `conflictlocal()` and `conflictother()` revsets return the
commits that are being merged, when there are conflicts. Also works
for conflicts caused by e.g. `hg graft`.
* `hg copy --forget` can be used to unmark a file as copied.
* The `format.revlog-compression` configuration entry now accept a list. The
first available option will be used. for example setting::
[format]
revlog-compression=zstd, zlib
Will use `zstd` compression for new repositories is available, and will
simply fall back to `zlib` if not.
* `hg debugmergestate` output is now templated, which may be useful
e.g. for IDEs that want to help the user resolve merge conflicts.
== New Experimental Features ==
* `hg copy` now supports a `--at-rev` argument to mark files as
copied in the specified commit. It only works with `--after` for
now (i.e., it's only useful for marking files copied using non-hg
`cp` as copied).
* Use `hg copy --forget --at-rev REV` to unmark already committed
copies.
== Bug Fixes ==
* Fix server exception when concurrent pushes delete the same bookmark
* Prevent pushes of divergent bookmarks (foo@remote)
* The push error "remote repository changed while pushing - please
try again" now only happens when a concurrent push changed related
heads (instead of when a concurrent pushed any revision).
== Backwards Compatibility Changes ==
* When `hg rebase` pauses for merge conflict resolution, the working
copy will no longer have the rebased node as a second parent. You
can use the new `conflictparents()` revset for finding the other
parent during a conflict.
* `hg rebase` now accepts repeated `--source` and `--base`
arguments. For example, `hg rebase --source 'A + B'` is equivalent
to `hg rebase --source A --source B`. This is a
backwards-incompatible change because it will break overriding an
alias `myrebase = rebase --source A` by `hg myrebase --source B`
(it will now rebase `(A + B)::` instead of `B::`).
* `hg recover` does not verify the validity of the whole repository
anymore. You can pass `--verify` or call `hg verify` if necessary.
* `hg debugmergestate` output format changed. Let us know if that is
causing you problems and we'll roll it back.
* Resolved merge conflicts are now cleared by `hg commit` even if the
working copy has no changes.
== Internal API Changes ==
* The deprecated `ui.progress()` has now been deleted. Please use
`ui.makeprogress()` instead.
* `hg.merge()` now takes a `ctx` instead of the previous `repo` and
`node` arguments.
* `hg.merge()` has lost its `abort` argument. Please call
`hg.abortmerge()` directly instead.
* `hg.merge()` has lost its `mergeforce` argument. It should have
only ever been called with the same value as the `force` argument.
* The `*others` argument of `cmdutil.check_incompatible_arguments()`
changed from being varargs argument to being a single collection.