Pierre-Yves David <pierre-yves.david@octobus.net> [Mon, 15 Mar 2021 16:37:11 +0100] rev 46767
perf-helper: add a new sampling revset based on anti-chain
See inline documentation for details.
Differential Revision: https://phab.mercurial-scm.org/D10222
Pierre-Yves David <pierre-yves.david@octobus.net> [Mon, 15 Mar 2021 16:35:54 +0100] rev 46766
perf-helper: add a small extension with revsets to select repository subset
Playing with discovery requires building interesting case. To do this we
need revsets to try to generate them. We start with a quite simple one. See
documentation for details.
Differential Revision: https://phab.mercurial-scm.org/D10221
Martin von Zweigbergk <martinvonz@google.com> [Tue, 26 Feb 2019 15:54:20 -0800] rev 46765
amend: mark commit obsolete after moving working copy
We were doing it this way:
1. move working copy (repo.setparents)
2. add obsmarkers (scmutil.cleanupnodes)
3. fix dirstate (dirstate.normal/drop)
Step 1 and 3 are closely related, so let's move them together. It
seems safest to create the obsmarkers last. This patch thus makes the
order 1, 3, 2.
Differential Revision: https://phab.mercurial-scm.org/D10197
Pierre-Yves David <pierre-yves.david@octobus.net> [Wed, 10 Mar 2021 06:03:01 +0100] rev 46764
pull: allow to specify multiple sources
I end up needing that on a regular basis and it turn out to be very simple to
implement. See documentation and test for details.
Differential Revision: https://phab.mercurial-scm.org/D10159
Pierre-Yves David <pierre-yves.david@octobus.net> [Wed, 10 Mar 2021 06:03:55 +0100] rev 46763
pull: pre-indent a part of the function
We are about to run it in a loop, so lets pre-indent it to clarify the actual
change in the next changesets.
Differential Revision: https://phab.mercurial-scm.org/D10158
Pierre-Yves David <pierre-yves.david@octobus.net> [Wed, 10 Mar 2021 05:54:27 +0100] rev 46762
command: clarify `postincoming` return and that return handling
The command should return None or a return code. The previous code was returning
boolean directly relying on the fact that `True → 1` and `False → 0`. This is a
good road to troubles, so lets be explicit about that return.
Differential Revision: https://phab.mercurial-scm.org/D10157
Pierre-Yves David <pierre-yves.david@octobus.net> [Wed, 10 Mar 2021 05:54:02 +0100] rev 46761
hg: make `clean` return consistent with the `update` function
The function return a boolean and is used as such. Lets be explicit about it.
Differential Revision: https://phab.mercurial-scm.org/D10156
Pierre-Yves David <pierre-yves.david@octobus.net> [Wed, 10 Mar 2021 16:28:46 +0100] rev 46760
tests: rename `test-pull` to `test-pull-network.t`
If is mostly about ssh and http interaction so lets avoid confusion with a
generic `hg pull` test
Differential Revision: https://phab.mercurial-scm.org/D10155
Pierre-Yves David <pierre-yves.david@octobus.net> [Wed, 10 Mar 2021 16:35:17 +0100] rev 46759
tests: move a test about update in test-pull-update.t
Differential Revision: https://phab.mercurial-scm.org/D10154
Martin von Zweigbergk <martinvonz@google.com> [Fri, 12 Mar 2021 09:15:40 -0800] rev 46758
split: close transaction in the unlikely event of a conflict while rebasing
`hg split` *should* never result in conflicts, but in case there are
bugs, we should at least commit the transaction so they can continue
the rebase. One of our users ran into the regression fixed by
D10120. They fixed the conflict and the tried to continue the rebase,
but it failed with "abort: cannot continue inconsistent rebase"
because the rebase state referred to commits written in a transaction
that was never committed.
Side note: `hg split` should probably turn off copy tracing to reduce
the impact of such bugs, and to speed it up as well. Copies made in
the rebased commits should still be respected because `hg rebase`
calls `copies.graftcopies()`.
Differential Revision: https://phab.mercurial-scm.org/D10164
Raphaël Gomès <rgomes@octobus.net> [Mon, 15 Mar 2021 13:05:00 +0100] rev 46757
rust-status: fix
issue6456 for the Rust implementation also
This implementation is being used by a few people now, so we need it fixed, even
though a big rewrite will be coming quite soon.
Differential Revision: https://phab.mercurial-scm.org/D10217
Joerg Sonnenberger <joerg@bec.de> [Tue, 16 Mar 2021 00:07:12 +0100] rev 46756
README: document requirement for builtin zstd
Differential Revision: https://phab.mercurial-scm.org/D10226
Simon Sapin <simon.sapin@octobus.net> [Mon, 08 Mar 2021 21:04:51 +0100] rev 46755
ci: Add a job testing with rhg installed as `hg`
This significantly increases test coverage of rhg, without duplicating
many tests that already exist.
The `cargo build` command being remove only compiled a `rhg` executable
(as shown by the preceding `cd` command) but since a previous patch
`run-tests.py --rhg` now takes care of that.
Differential Revision: https://phab.mercurial-scm.org/D10196
Pierre-Yves David <pierre-yves.david@octobus.net> [Wed, 10 Mar 2021 13:54:46 +0100] rev 46754
run-test: install rhg if --rhg is passed
Before this, --rhg was only working with --local.
Differential Revision: https://phab.mercurial-scm.org/D10195
Pierre-Yves David <pierre-yves.david@octobus.net> [Wed, 10 Mar 2021 13:54:13 +0100] rev 46753
makefile: add a install option
This gives and easy way to install rhg that we can use in `run-test.py` in the
next changesets.
Differential Revision: https://phab.mercurial-scm.org/D10194
Pierre-Yves David <pierre-yves.david@octobus.net> [Wed, 10 Mar 2021 13:54:00 +0100] rev 46752
makefile: add a build-rhg option
This gives an easy action to build the rhg-binary. This will be useful for the
`install-rhg` action in the next changeset.
Differential Revision: https://phab.mercurial-scm.org/D10193
Pierre-Yves David <pierre-yves.david@octobus.net> [Wed, 10 Mar 2021 13:53:47 +0100] rev 46751
makefile: add a install-chg option
This is done as a gratuitous improvement on the way to add makefile entry to
build and install rhg.
It seems saner to have equivalent entry for chg too.
Differential Revision: https://phab.mercurial-scm.org/D10192
Pierre-Yves David <pierre-yves.david@octobus.net> [Wed, 10 Mar 2021 13:53:31 +0100] rev 46750
makefile: add a build-chg option
This is done as a gratuitous improvement on the way to add makefile entry to
build and install rhg.
It seems saner to have equivalent entry for chg too.
Differential Revision: https://phab.mercurial-scm.org/D10191
Simon Sapin <simon.sapin@octobus.net> [Fri, 12 Mar 2021 23:04:56 +0100] rev 46749
tests: Disable for rhg remaining tests that fail in that mode
These cases are in features not yet implemented by rhg
for which triggering a fallback is not practical.
Disabling some tests allows us to reach passing CI and catch
any future regression in the rest of the tests.
Differential Revision: https://phab.mercurial-scm.org/D10190
Simon Sapin <simon.sapin@octobus.net> [Fri, 12 Mar 2021 21:44:07 +0100] rev 46748
rhg: Remove `rhg.fallback-executable=hg` default configuration
When `rhg.on-unsupported` is configured to `fallback` and an unsupported
feature is encountered, the previous default was to look for an `hg`
executable in `$PATH`.
This default was fragile since it was easy to end up accidentally using
an older version of Mercurial installed system-wide,
when a local (perhaps patched) installation was intended.
Instead, it is now an error to have `rhg.on-unsupported=fallback`
without also configuring an explicit path or the fallback executable.
Differential Revision: https://phab.mercurial-scm.org/D10189
Simon Sapin <simon.sapin@octobus.net> [Tue, 09 Mar 2021 09:17:24 +0100] rev 46747
rhg: Return an error code for `rhg config Section.idontexist`
This is what Python-based hg does.
Differential Revision: https://phab.mercurial-scm.org/D10145
Simon Sapin <simon.sapin@octobus.net> [Mon, 08 Mar 2021 20:04:20 +0100] rev 46746
rhg: Fall back to Python for --version
Clap has some built-in support for printing something on --version,
but it looks different than what Pytho-based hg does.
Also, at the moment we’re not giving version numbers to rhg
separate from the Mercurial version.
Differential Revision: https://phab.mercurial-scm.org/D10144
Simon Sapin <simon.sapin@octobus.net> [Mon, 08 Mar 2021 19:25:33 +0100] rev 46745
rhg: Exit with an error code if `files` finds nothing
This matches the behavior of Python-based hg.
Differential Revision: https://phab.mercurial-scm.org/D10143
Simon Sapin <simon.sapin@octobus.net> [Wed, 03 Mar 2021 16:40:03 +0100] rev 46744
rhg: `cat` command: print error messages for missing files
And exit with an error code if no file was matched.
This matches the behavior of Python-based hg.
Differential Revision: https://phab.mercurial-scm.org/D10142
Simon Sapin <simon.sapin@octobus.net> [Mon, 08 Mar 2021 19:07:29 +0100] rev 46743
rhg: Silently ignore missing files in config %include
… instead of aborting with an error message.
This is what Python-based hg does.
Differential Revision: https://phab.mercurial-scm.org/D10141
Simon Sapin <simon.sapin@octobus.net> [Mon, 08 Mar 2021 15:35:32 +0100] rev 46742
rhg: Add support for environment variables in config include paths
Some tests rely on this.
Differential Revision: https://phab.mercurial-scm.org/D10140
Simon Sapin <simon.sapin@octobus.net> [Mon, 08 Mar 2021 08:55:48 +0100] rev 46741
rhg: Add support for the HGRCSKIPREPO environment variable
It’s easy enough and affects tests, well, that test that feature.
Differential Revision: https://phab.mercurial-scm.org/D10139
Simon Sapin <simon.sapin@octobus.net> [Mon, 08 Mar 2021 08:35:43 +0100] rev 46740
rhg: Don’t make repository path absolute too early
Some error messages want to include a relative path,
which affects the output of some tests.
Differential Revision: https://phab.mercurial-scm.org/D10138
Simon Sapin <simon.sapin@octobus.net> [Fri, 05 Mar 2021 13:28:49 +0100] rev 46739
rhg: Fall back to Python if ui.relative-paths is configured
This feature is not supported yet, and affects the output of some tests.
Differential Revision: https://phab.mercurial-scm.org/D10137
Pierre-Yves David <pierre-yves.david@octobus.net> [Mon, 08 Mar 2021 17:32:22 +0100] rev 46738
tests: clarify some missing output in test-merge-subrepos
This makes the test behavior clearer, especially, why the output are
inconsistent.
Differential Revision: https://phab.mercurial-scm.org/D10136
Simon Sapin <simon.sapin@octobus.net> [Fri, 05 Mar 2021 11:39:21 +0100] rev 46737
tests: Adapt expected output for minor differences with rhg
Differential Revision: https://phab.mercurial-scm.org/D10135
Simon Sapin <simon.sapin@octobus.net> [Fri, 05 Mar 2021 11:21:15 +0100] rev 46736
rhg: Add support for --cwd
This affect the meaning of relative paths in `--repository`,
which are resolved "early" by rhg in order to load config
which is needed before fallback to Python is considered.
An incorrect path could cause errors when loading a non-existent repo,
leading to failing tests even when fallback is enabled.
Differential Revision: https://phab.mercurial-scm.org/D10134
Simon Sapin <simon.sapin@octobus.net> [Thu, 04 Mar 2021 13:30:20 +0100] rev 46735
rhg: Align with Python on some more error messages
Differences can cause some tests to fail
Differential Revision: https://phab.mercurial-scm.org/D10133
Simon Sapin <simon.sapin@octobus.net> [Fri, 12 Mar 2021 22:38:40 +0100] rev 46734
rhg: Add an allow-list of ignored extensions
Because rhg doesn’t know how a Python extension would affect
behavior it implements in Rust, when an unsupported extension
is enabled it conservatively falls back to Python-based hg.
However many users will have unsupported extensions enabled in practice.
Maybe they don’t actually affect rhg behavior, but we don’t know.
This adds a `rhg.ignored-extensions` configuration that lets
users list extensions that rhg can safely ignore and proceed even
if they’re not supported in Rust.
Differential Revision: https://phab.mercurial-scm.org/D10188
Simon Sapin <simon.sapin@octobus.net> [Thu, 04 Mar 2021 10:58:43 +0100] rev 46733
rhg: Fall back to Python if unsupported extensions are enabled
Extensions might affect behavior in ways we can’t anticipate,
so just ignoring them is not correct.
Later we’ll add opt-in configuration to ignore specific extensions.
Differential Revision: https://phab.mercurial-scm.org/D10112
Simon Sapin <simon.sapin@octobus.net> [Wed, 03 Mar 2021 20:02:07 +0100] rev 46732
rhg: Sort config files when adding a directory
For example in `/etc/mercurial/hgrc.d/` or with `HGRCPATH=some-directory`.
Previously files where parsed in the order returned by the filesystem,
which is undefined. But order is significant when multiple files
define the same configuration key: the "last" one wins.
Differential Revision: https://phab.mercurial-scm.org/D10111
Simon Sapin <simon.sapin@octobus.net> [Wed, 03 Mar 2021 19:47:48 +0100] rev 46731
rhg: Align config file parse error formatting with Python
Differences can cause tests to fail
Differential Revision: https://phab.mercurial-scm.org/D10110
Simon Sapin <simon.sapin@octobus.net> [Wed, 03 Mar 2021 19:08:27 +0100] rev 46730
rhg: Fall back to Python for bundle repositories
rhg does not support bundles at all, yet.
Differential Revision: https://phab.mercurial-scm.org/D10102
Simon Sapin <simon.sapin@octobus.net> [Wed, 03 Mar 2021 19:02:06 +0100] rev 46729
rhg: Fall back to Python on --repository with an URL
A low-hanging fruit to improve on this would be to properly parse
and handle `file:` URLs. But other Python-based hg supports
some other URL schemes for features that rhg does not support yet.
Differential Revision: https://phab.mercurial-scm.org/D10101
Simon Sapin <simon.sapin@octobus.net> [Wed, 03 Mar 2021 18:43:05 +0100] rev 46728
rhg: Print non-absolutized path in "repository {} not found" errors
… like Python does. Differences in ouput can cause tests to fail.
Differential Revision: https://phab.mercurial-scm.org/D10100
Simon Sapin <simon.sapin@octobus.net> [Wed, 03 Mar 2021 18:40:17 +0100] rev 46727
rhg: Align "malformed --config" error message with Python
Differences in error message formatting can cause tests to fail.
Differential Revision: https://phab.mercurial-scm.org/D10099
Simon Sapin <simon.sapin@octobus.net> [Wed, 03 Mar 2021 18:38:22 +0100] rev 46726
rhg: Fall back to Python on unsupported `rhg config <section>`
Printing an entire section (as opposed to a single value with
`rhg config foo.bar`) is not supported yet in Rust only.
Differential Revision: https://phab.mercurial-scm.org/D10098
Simon Sapin <simon.sapin@octobus.net> [Tue, 02 Mar 2021 23:18:23 +0100] rev 46725
rhg: Fall back to Python for unsupported revset syntax
rhg only supports a small subset of the syntax.
On parse error, this gives Python a chance instead of aborting immediately.
Differential Revision: https://phab.mercurial-scm.org/D10097
Simon Sapin <simon.sapin@octobus.net> [Wed, 03 Mar 2021 18:00:54 +0100] rev 46724
tests: Add `rhg` and `no-rhg` for #require and #if in .t files
With this we can allow some small differences in expected output
between Rust-based and Python-based code paths.
Differential Revision: https://phab.mercurial-scm.org/D10096
Simon Sapin <simon.sapin@octobus.net> [Tue, 02 Mar 2021 22:42:10 +0100] rev 46723
tests: Enable rhg fallback to Python by default in tests
This premise of `run-tests.py --rhg`: fallback should make
`rhg` behave the same as `hg`, except faster in some cases.
To test run the whole test suite with installed `rhg` as `hg`
and with fallback enabled.
Differential Revision: https://phab.mercurial-scm.org/D10187
Simon Sapin <simon.sapin@octobus.net> [Fri, 12 Mar 2021 18:45:32 +0100] rev 46722
rhg: Add environment variables for fallback configuration
For the `rust-tests.py --rhg` we want every `hg` command in tests
to run `rhg` with fallback enabled, but other methods of setting
configuration are limited or disruptive as discussed in code comment.
Differential Revision: https://phab.mercurial-scm.org/D10186
Pierre-Yves David <pierre-yves.david@octobus.net> [Thu, 11 Mar 2021 11:22:54 +0100] rev 46721
paths: add a `*` special path to define default sub option
Differential Revision: https://phab.mercurial-scm.org/D10163
Pierre-Yves David <pierre-yves.david@octobus.net> [Thu, 11 Mar 2021 17:26:49 +0100] rev 46720
ui: pass a `ui` object to `paths.getpath`
I want to introduce more path's suboption and make it possible to use default
value for them. Processing theses sub-options might result in warnings. We
need a `ui` object to issue such warnings.
To make things simpler, we add an helper method on the `ui` object.
Differential Revision: https://phab.mercurial-scm.org/D10162
Raphaël Gomès <rgomes@octobus.net> [Mon, 15 Mar 2021 10:57:02 +0100] rev 46719
configitems: add TODOs blocking the move out of experimental for revlogv2
These are the todos so far, we probably will have more, but we might as well
list them while they're fresh in our minds.
Differential Revision: https://phab.mercurial-scm.org/D10216
Raphaël Gomès <rgomes@octobus.net> [Fri, 19 Feb 2021 11:24:50 +0100] rev 46718
sidedata-exchange: rewrite sidedata on-the-fly whenever possible
When a A exchanges with B, the difference of their supported sidedata categories
is made, and the responsibility is always with the client to generated it:
- If A pushes to B and B requires category `foo` that A does not have, A
will need to generate it when sending it to B.
- If A pulls from B and A needs category `foo`, it will generate `foo`
before the end of the transaction.
- Any category that is not required is removed. If peers are not compatible,
abort.
It is forbidden to rewrite sidedata for a rev that already has sidedata, since
that would introduce unreachable (garbage) data in the data file, something
we're not prepared for yet.
Differential Revision: https://phab.mercurial-scm.org/D10032
Raphaël Gomès <rgomes@octobus.net> [Mon, 15 Feb 2021 11:08:28 +0100] rev 46717
revlog-index: add `replace_sidedata_info` method
During a `pull` operation where the server does not provide sidedata, the client
that requires it should generate them on-the-fly. In the generic case, we need
to wait for the changelog + manifests + filelogs to be added, since we don't
know what the sidedata computers might need: this means rewriting the sidedata
of index entries from within the pull transaction (and no further back) right
after we've added them.
Both Python and C implementations only allow for rewriting the sidedata offset
and length for revs within the transaction where they were created.
Differential Revision: https://phab.mercurial-scm.org/D10031
Raphaël Gomès <rgomes@octobus.net> [Fri, 19 Feb 2021 11:04:17 +0100] rev 46716
revlogv2: temporarily forbid inline revlogs
See inline comments. I plan to fix the underlying issue before revlogv2 is
stabilized.
Differential Revision: https://phab.mercurial-scm.org/D10030
Raphaël Gomès <rgomes@octobus.net> [Fri, 19 Feb 2021 11:15:42 +0100] rev 46715
changegroupv4: add sidedata helpers
These helpers carry the information and computers needed to rewrite sidedata
when generating/applying patches. We will be making use of them soon.
Differential Revision: https://phab.mercurial-scm.org/D10029
Raphaël Gomès <rgomes@octobus.net> [Wed, 10 Mar 2021 19:33:18 +0100] rev 46714
revlog: add attribute on revlogs that specifies its kind
The sidedata logic needs to check whether the revlog it's working on is a
changelog, a manifest or a filelog.
Furthermore, future versions of the revlog format will most likely see a split
between the three types (i.e. they will store different information), so having
this will be useful for other future endeavors as well.
Differential Revision: https://phab.mercurial-scm.org/D10151
Raphaël Gomès <rgomes@octobus.net> [Fri, 19 Feb 2021 10:53:27 +0100] rev 46713
sidedata-exchange: add `wanted_sidedata` and `sidedata_computers` to repos
Each repo will advertise the sidedata categories it requires (categories being
unique and canonical), and have a set of "computers", functions to generate
sidedata from `(repo, revlog, rev, previous_sidedata)`, for a given category.
The set of computers can be a superset of the set of the wanted categories, but
not smaller: repos are expected to be coherent in their handling of sidedata.
Differential Revision: https://phab.mercurial-scm.org/D10028
Raphaël Gomès <rgomes@octobus.net> [Thu, 18 Feb 2021 18:18:35 +0100] rev 46712
delta: add sidedata field to revision delta
When emitting revision delta, we need to also emit the sidedata information just
added in the revlogv2 format if appropriate.
Differential Revision: https://phab.mercurial-scm.org/D10027