This is an archive of the discontinued Mercurial Phabricator instance.

Differential D6557

relnotes: added description about statemod._statecheck
ClosedPublic

Authored by taapas1128 on Jun 20 2019, 2:56 PM.

Details

Reviewers
martinvonz
Group Reviewers
hg-reviewers
Commits
rHGdf5f674050b7: relnotes: added description about statemod._statecheck

Diff Detail

Repository
rHG Mercurial
Lint
Lint Skipped
Unit
Unit Tests Skipped

Event Timeline

taapas1128 created this revision.Jun 20 2019, 2:56 PM
martinvonz added a subscriber: martinvonz.Jun 27 2019, 1:08 PM
martinvonz added inline comments.
relnotes/next
63–69

The "private" (underscore-prefixed) members should not be mentioned here. Extension developers should not be accessing them, so they don't need to know about them. What they need to know is that if they had been adding operations to cmdutil.unfinishedstates, cmdutil.afterresolvestates, or cmdutil.STATES before, they now need to use state.addunfinished().

I now see that state.addunfinished() has **kwargs, which means they will have to look at the _statecheck class after all. Perhaps we should copy the arguments over from the class to the function (and remove the defaults from the class) so it's easier to find them?

taapas1128 added inline comments.Jun 27 2019, 3:24 PM
relnotes/next
63–69

@martinvonz But that would destroy the integrity of statecheck class and addunfinished is just below statecheck class so one can easily look at statecheck if definitions for various flags are to be known. Still if you want I can do that.

taapas1128 updated this revision to Diff 15690.Jun 27 2019, 6:02 PM
taapas1128 marked an inline comment as done.Jun 27 2019, 6:03 PM
martinvonz accepted this revision.Jun 27 2019, 6:06 PM
martinvonz added inline comments.
relnotes/next
65–66

I'll drop the last sentence in flight since D6583 has landed.

This revision is now accepted and ready to land.Jun 27 2019, 6:06 PM
taapas1128 added inline comments.Jun 27 2019, 6:08 PM
relnotes/next
65–66

thanks!

Revision Contents
Changeset List

PathPackages
Mrelnotes/next (14 lines)

Diff 15620

relnotes/next

* Matchers are no longer iterable. Use `match.files()` instead. * Matchers are no longer iterable. Use `match.files()` instead.
* `match.visitdir()` and `match.visitchildrenset()` now expect the * `match.visitdir()` and `match.visitchildrenset()` now expect the
empty string instead of '.' to indicate the root directory. empty string instead of '.' to indicate the root directory.
* `util.dirs()` and `util.finddirs()` now include an entry for the * `util.dirs()` and `util.finddirs()` now include an entry for the
root directory (empty string). root directory (empty string).
* New API to manage unfinished operations: Earlier there were distinct APIs
which dealt with unfinished states and separate lists maintaining them.
eg:`cmdutil.afterresolvestates`, `cmdutil.unfinishedstates` and
`cmdutil.STATES`. Now these have been unified to a single
API which handles the various states and their utilities.This API
has been added to `state.py` and is form of class called `_statecheck`.
To register a new state for an operation you need to add a `_statecheck`
object with necessary information to `_unfinishedstates` list.
The method of registration of a `_statecheck` object into `_unfinishedstates`
martinvonzUnsubmitted
Not Done

I'll drop the last sentence in flight since D6583 has landed.

martinvonz: I'll drop the last sentence in flight since D6583 has landed.
taapas1128AuthorUnsubmitted
Done

thanks!

taapas1128: thanks!
is not direct but through `addunfinished()`. The data needed for various
operations to be registered under `_unfinishedstates` list is stated in
`_statecheck` class.
martinvonzUnsubmitted
Done

The "private" (underscore-prefixed) members should not be mentioned here. Extension developers should not be accessing them, so they don't need to know about them. What they need to know is that if they had been adding operations to cmdutil.unfinishedstates, cmdutil.afterresolvestates, or cmdutil.STATES before, they now need to use state.addunfinished().

I now see that state.addunfinished() has **kwargs, which means they will have to look at the _statecheck class after all. Perhaps we should copy the arguments over from the class to the function (and remove the defaults from the class) so it's easier to find them?

martinvonz: The "private" (underscore-prefixed) members should not be mentioned here. Extension developers…
taapas1128AuthorUnsubmitted
Not Done

@martinvonz But that would destroy the integrity of statecheck class and addunfinished is just below statecheck class so one can easily look at statecheck if definitions for various flags are to be known. Still if you want I can do that.

taapas1128: @martinvonz But that would destroy the integrity of `statecheck` class and `addunfinished` is…