Pierre-Yves David <pierre-yves.david@octobus.net> [Tue, 28 Sep 2021 18:57:20 +0200] rev 48063
dirstate: drop some safety assert in largefile
The code involved in `set_possibly_dirty` is now simpler and safe to use even
in the cases that the assert covered. So we can drop this assert.
It was the last user of `merged_removed` and `from_p2_removed`.
Differential Revision: https://phab.mercurial-scm.org/D11511
Pierre-Yves David <pierre-yves.david@octobus.net> [Tue, 28 Sep 2021 18:29:57 +0200] rev 48062
dirstate: drop unused condition in `from_p2`
This conditional was added (by me) tentatively because "it seemed more
correct", but it is not used anywhere yet, and it is missing from the C and the
Rust implementation. So it seems more consistent to drop it for now.
This effectively backout
f94cc63df859c
Differential Revision: https://phab.mercurial-scm.org/D11510
Pierre-Yves David <pierre-yves.david@octobus.net> [Tue, 28 Sep 2021 20:05:37 +0200] rev 48061
dirstate: drop all logic around the "non-normal" sets
The dirstate has a lot of code to compute a set of all "non-normal" and
"from_other_parent" entries.
This is all used in one, unique, location, when `setparent` is called and moved
from a merge to a non merge. At that time, any "merge related" information has
to be dropped. This is mostly useful for command like `graft` or `shelve` that
move to a single-parent state -before- the commit. Otherwise the commit will
already have removed all traces of the merge information in the dirstate (e.g.
for a regular merges).
The bookkeeping for these sets is quite invasive. And it seems simpler to just
drop it and do the full computation in the single location where we actually
use it (since we have to do the computation at least once anyway).
This simplify the code a lot, and clarify why this kind of computation is
needed.
The possible drawback compared to the previous code are:
- if the operation happens in a loop, we will end up doing it multiple time,
- the C code to detect entry of interest have been dropped, for now. It will be
re-introduced later, with a processing code directly in C for even faster
operation.
Differential Revision: https://phab.mercurial-scm.org/D11507
Pierre-Yves David <pierre-yves.david@octobus.net> [Wed, 22 Sep 2021 17:46:29 +0200] rev 48060
dirstate: use a new `drop_merge_data` in `setparent`
What is happening in this `setparent` loop is that we remove all `merge`
related information when the dirstate is moved out of a `merge` situation.
So instead of shuffling state to get them where we want, we simply add a method
on the DirstateItem to do drop the information we want dropped.
Differential Revision: https://phab.mercurial-scm.org/D11506
Pierre-Yves David <pierre-yves.david@octobus.net> [Wed, 22 Sep 2021 15:17:12 +0200] rev 48059
dirstate: move parent state handling in the dirstatemap
This involves dirstatemap data mostly. Moving this one level down will remove
the needs for the dirstatemap to expose some of its internals.
This will help us to simplify more code further.
Differential Revision: https://phab.mercurial-scm.org/D11505
Pierre-Yves David <pierre-yves.david@octobus.net> [Wed, 22 Sep 2021 09:46:37 +0200] rev 48058
dirstate: stop checking for path collision when adjusting parents
This was already checked at a earlier point when adding the file.
Differential Revision: https://phab.mercurial-scm.org/D11504
Pierre-Yves David <pierre-yves.david@octobus.net> [Wed, 22 Sep 2021 15:08:47 +0200] rev 48057
dirstate: drop the `_updatedfiles` set
This is a lot of book keeping for something that was only used to clear
ambiguous time. Since this is no no longer in use, we can drop it.
Differential Revision: https://phab.mercurial-scm.org/D11503
Pierre-Yves David <pierre-yves.david@octobus.net> [Wed, 22 Sep 2021 15:23:03 +0200] rev 48056
dirstate: drop the `clearambiguoustimes` method for the map
This is no longer called anywhere.
Differential Revision: https://phab.mercurial-scm.org/D11502
Pierre-Yves David <pierre-yves.david@octobus.net> [Wed, 22 Sep 2021 14:54:42 +0200] rev 48055
dirstate: simplify the ambiguity clearing at write time
The serialization function is already doing this, so we don't need to do it
manually. We just need to propagate the right definition of "now".
Differential Revision: https://phab.mercurial-scm.org/D11501
Martin von Zweigbergk <martinvonz@google.com> [Tue, 28 Sep 2021 09:32:24 -0700] rev 48054
histedit: use more specific exceptions for more detailed exit codes
Differential Revision: https://phab.mercurial-scm.org/D11509
Martin von Zweigbergk <martinvonz@google.com> [Tue, 28 Sep 2021 09:25:05 -0700] rev 48053
histedit: remove redundant checks for unfinished histedit state
Both text-based and curses-based histedit already check for unfinished
operations (not just unfinished histedit), so there's no need to check
specifically for unfinished histedit.
Differential Revision: https://phab.mercurial-scm.org/D11508
Simon Sapin <simon.sapin@octobus.net> [Tue, 28 Sep 2021 13:43:14 +0200] rev 48052
dirstate: Appease pytype
test-check-pytype.t was failing since
98c0408324e6:
File "/home/simon/projects/hg/mercurial/dirstatemap.py", line 572, in
addfile: unsupported operand type(s) for &: 'size: None' and
'rangemask: int' [unsupported-operands]
No attribute '__and__' on 'size: None' or '__rand__' on 'rangemask: int'
File "/home/simon/projects/hg/mercurial/dirstatemap.py", line 573, in
addfile: unsupported operand type(s) for &: 'mtime: None' and
'rangemask: int' [unsupported-operands]
No attribute '__and__' on 'mtime: None' or '__rand__' on 'rangemask: int'
`None` is the default value of the `size` and `mtime` parameters of the
`addfile` method. However, the relevant lines are only used in a code path
where those defaults are never used. These `size` and `mtime` are passed
to `DirstateItem.new_normal` which (in the C implementation) calls
`dirstate_item_new_normal` which uses:
PyArg_ParseTuple(args, "iii", &mode, &size, &mtime)
So `None` values would cause an exception to be raised anyway.
The new `assert`s only move that exception earlier, and informs pytype
that we expect `None` to never happen in this code path.
Differential Revision: https://phab.mercurial-scm.org/D11500
Simon Sapin <simon.sapin@octobus.net> [Thu, 23 Sep 2021 18:29:40 +0200] rev 48051
dirstate: Pass the final DirstateItem to _rustmap.addfile()
Now that the Python DirstateItem class wraps a Rust DirstateEntry value,
use that value directly instead of converting through v1 data + 5 booleans.
Also remove propogating the return value. None of the callers look at it,
and it is always None.
Differential Revision: https://phab.mercurial-scm.org/D11494
Simon Sapin <simon.sapin@octobus.net> [Thu, 23 Sep 2021 15:36:43 +0200] rev 48050
dirstate: Replace dropfile with drop_item_and_copy_source
Those removing a DirstateItem and a copy source are always done together
Differential Revision: https://phab.mercurial-scm.org/D11493
Simon Sapin <simon.sapin@octobus.net> [Thu, 23 Sep 2021 15:29:38 +0200] rev 48049
rust: Remove some obsolete doc-comments
About parameters that have been removed or replaced
Differential Revision: https://phab.mercurial-scm.org/D11492
Simon Sapin <simon.sapin@octobus.net> [Thu, 23 Sep 2021 15:26:33 +0200] rev 48048
dirstate: Remove return boolean from dirstatemap.dropfile
None of the remaining callers use it.
Differential Revision: https://phab.mercurial-scm.org/D11491
Simon Sapin <simon.sapin@octobus.net> [Wed, 22 Sep 2021 18:56:58 +0200] rev 48047
dirstate: Propagate dirstate-v2 parse errors from set_dirstate_item
… so that Python sees a proper ValueError instead of only
`SystemError: Rust panic`
Differential Revision: https://phab.mercurial-scm.org/D11489
Simon Sapin <simon.sapin@octobus.net> [Wed, 22 Sep 2021 18:42:00 +0200] rev 48046
dirstate: Don’t drop unrelated data in DirstateMap::set_entry
For example, copy source are handled separately. Removing it goes through
the `copy_map_remove` method (exposed to Python as `.copymap.pop()`)
Differential Revision: https://phab.mercurial-scm.org/D11488
Simon Sapin <simon.sapin@octobus.net> [Wed, 22 Sep 2021 18:21:58 +0200] rev 48045
dirstate: Skip no-op conversion in Rust DirstateMap::set_v1
Now that the `DirstateItem` python class is implemented in Rust containing
a `DirstateEntry` value, use that value directly instead of reconstructing
it from v1 data.
Also rename from `set_v1` since dirstate-v1 data is not used anymore.
Differential Revision: https://phab.mercurial-scm.org/D11487
Simon Sapin <simon.sapin@octobus.net> [Wed, 22 Sep 2021 11:33:29 +0200] rev 48044
dirstate: Use the Rust implementation of DirstateItem when Rust is enabled
… instead of the C implementation, with C/Rust conversions at the FFI boundary
Differential Revision: https://phab.mercurial-scm.org/D11486
Simon Sapin <simon.sapin@octobus.net> [Wed, 22 Sep 2021 11:28:52 +0200] rev 48043
rust: Add Python bindings for DirstateEntry as rustext.dirstate.DirstateItem
Differential Revision: https://phab.mercurial-scm.org/D11485
Simon Sapin <simon.sapin@octobus.net> [Mon, 20 Sep 2021 19:18:21 +0200] rev 48042
rust: Align DirstateEntry internals with Python/C DirstateItem
This propagate to this Rust struct the similar change that was made recently
to the Python classe and C struct. Namely, instead of storing a four-valued
`state` field we now store seven (bit-packed) booleans that give lower-level
information.
Additionally, the marker values -1 and -2 for mtime and size should not
be used internally anymore. They are replaced by some combinations of booleans
For now, all uses of of `DirstateEntry` still use the compatibility APIs
with `state` and marker values. Later the Rust API for DirstateMap
will be increasingly updated to the new style.
Also change the expected result of the test_non_normal_other_parent_entries
unit test. Only a `DirstateEntry` with `size == -2 && mtime != -1` is
affected, but this case never occurs outside of unit tests.
`size == -2` was the marker value for "from other parent" entries,
where no meaningful mtime is stored.
Differential Revision: https://phab.mercurial-scm.org/D11484
Raphaël Gomès <rgomes@octobus.net> [Tue, 28 Sep 2021 09:40:57 +0200] rev 48041
branching: merge stable into default
Pierre-Yves David <pierre-yves.david@octobus.net> [Tue, 21 Sep 2021 18:18:56 +0200] rev 48040
manifestlog: also monitor `00manifest.n` when applicable
This let the locarepo's file cache detect outdated nodemap docket and reload the
manifestlog after `localrepo.invalidate` when applicable.
The same problem than
issue6554 could affect the Manifest too.
Differential Revision: https://phab.mercurial-scm.org/D11483