rewriteutil: allow rewriting merge commits (issue4561) This patch simply allows rewriteutil.rewrite() to work with commits with multiple parents (i.e. merges). That function is used in such commands as fold, metaedit, touch, rewind. The issue 4561 is marked as easy, the limitation is called unnecessary, no tests fail after this change. What can go wrong.

tests: show what happens when trying to hg touch a merge commit

evolve: further clarify that update is performed only when requested Text further modified by Pierre-Yves David and Anton Shestakov.

touch: let's not use util.acceptintervention() as it's not required

branching: merge with stable

evolve: use right value for branch name when finding branch heads subbranch already formatted as "branchname:topicname", again appending it with ":topicname" doesn't not make sense. It's a little bit surprising that no tests fails though.

evolve: fix confusion in branch heads checking logic when topic in play To provide some context, when topics are in play the branchmap cache we store contains the branch info of a rev as "branch:topic" format IIUC. Assuming that is right, now in present code we don't actually cover this part that "when looking for branch heads where we also have active topic we should look for branch='branch_name:topic' instead". And we get wrong branch heads as a result. This patch make sure that we pass right candidate to find branch heads using branchmap.branchheads() by overriding the localrepo.branchheads() Changes in test file reflect the fixed behavior.

topic: add tests to demonstrate topic confuses the branchhead checking logic While topics are in play, we store the branchheads (which has a topic) in "branchname:topicname" format. After digging into it I found that even in the case when we should have branch heads for "bname:tname" we get heads for "bname". The tests output reflect the confusion in branch head checking logic. Next patch will be fixing the problem.

evolve: fix the inconsistent behaviour of prune (issue6137) Let's not update to any revision when working directory parent is not related to the revision being pruned. Changes in test file demonstrate the fixed behaviour.

prune: add tests to demonstrate issue6137 Here we can see that prune updates off to the parent revision even when the pruned revision wasn't related with the working directory parent. A follow-up patch will fix this.

compat: fix `setupevolveunfinished` for upstream

prune: update to the successor of wdir also with --pair/--biject (issue6142) When prune is used with --pair flag, we can also update to the successor of working directory parent. No need to check len(sucs) or len(precs) here because there's a check for that earlier in the code (it's a requirement of biject). The tests are now demonstrate the correct behavior: when rev 14 was pruned with 12 as its successor, the bookmark that was on 14 was moved to 12. That bookmark was also activated (even before this patch).

tests: demonstrate prune --pair not moving bookmark correctly After `mkcommit n2` line the bookmark is on the correct changeset, but when we prune --pair the two newly created changesets (revs 13 and 14), the bookmark gets moved to their ancestor (rev 0). Instead, it should've moved to the last of their successors (rev 12).

oops: backed out changeset 7ac40b4ea24c Anton requested some changes on it.

compat: fix `setupevolveunfinished` for upstream

evolve: fix the code flow pattern of solving obswdir par and troubled revs Now we will go to _handlenotrouble() (which prints messages about no revs to solve) only when there is no troubled revs and working dir parent is not obsolete. This change also saves us from an issue which was about looking into the revset (smartset contains troubled revs to solve) when a rev from the revset gets hidden. This happens in the case when our wdir parent is obsolete. After resolving obswdir parent we were looking into the revset to check if there is any troubled revs to solve but we should have performed this check before performing the obswdir resolution. Changes in test file reflect this fixed behaviour.

evolve: refactor for consistent behavior of evolve when wdp is obsolete This patch make sure that when working directory parent is obsolete `hg evolve` and `hg evolve --all` don't behave differently.

evolve: backout 3027005c42c3 to reintroduce a bug for right fix This patch backout 3027005c42c3 as it was accepted by mistake while it was being "in-discussion" state.

pick: register pickstate as an unfinished state This way pickstate file will indicate that unfinished pick command needs to be dealt with (--continue or --abort) before modifying the repo. Otherwise it would be e.g. possible to commit during an interrupted pick and that's not expected.

pick: rename variable for unfinishedstates

pick: actually delete pickstate if --abort is given Makes pick to be, uh, actually aborted.

evolve: orphans that evolve into nothing don't need successors (issue5967) When continuing to solve an orphan that created no changes (i.e. clean wdir), _completeorphan() used to create an obsmarker that said that the result of that orphan evolution is the currently checked out changeset. That's not a correct obsmarker, because all of the orphan's changes were dropped and so it had no effect on the currently checked out changeset. This is an issue that has only existed when --continu'ing evolve, that's why the fix touches _completeorphan(), but not _solveunstable(). This fix is adapted from a similar "if node is None" block in _finalizerelocate().

tests: demonstrate obsmarker creation after discarding conflicting changes Continued evolve creates an incorrect obsmarker that says 2 is a successor of 1. It's incorrect because 1 was dropped as it created no changes to commit (after conflict resolution that discarded its changes). If evolve does the same thing in one go (e.g. just by using --tool :local and without subsequent need to continue) the obsmarker is correct.

pick: remove transaction on the whole command (issue6037) At its core, pick is a pretty straightforward and well-behaving command, it uses functions already in core hg, it checks that wdir is clean and that changeset to pick is not public, it checks if there happen to be merge conflicts and can be --continue'd later, etc. It is very similar to graft in core (it also uses mergemod.graft function), but it obsoletes the original changeset. However, graft does not experience this incorrect behavior from issue 6037. What happens in the test case for this issue when we pick a revision that touches both "a" and "b": mergemod.graft() takes the original changeset and tries to apply it to the wdir, which results in "b" being marked as newly added and ready to be committed, "a" updated with the new content and being marked as modified, but "a" also has conflicts. Pick correctly notices this and saves its state before asking for user intervention. So far so good. However, when the command raises InterventionRequired to print a user-facing message and exit while being wrapped in repo.transaction() context manager, the latter partially undoes what mergemod.graft() did: it unmarks "b" as added. And when user continues pick, "b" is therefore not tracked and is not included in the resulting commit. The transaction is not useful here, because it doesn't touch wdir (it's still dirty), it doesn't remove pickstate (and other commands will refuse to work until pick --abort or --continue), it just makes "b" untracked. The solution is to use repo.transaction() only to wrap code that writes data to hg store in the final stages of the command after all checks have passed and is not expected to fail on trivial cases like merge conflicts. For example, committing the picked changeset. But since pick uses repo.commit() for that, and because that function already uses a transaction, wrapping it in another transaction doesn't make sense.

tests: demonstrate hg pick forgetting files after conflicts This test currently passes to show that pick is behaving incorrectly.

packaging: follow hg's supported python version (>= 2.7) "Mercurial 4.3 and newer require Python 2.7." (From https://www.mercurial-scm.org/wiki/SupportedPythonVersions) Also add X- prefix, because that's the correct form, apparently. This line can also be removed in future, since "When Debian supported multiple Python versions, X-Python-Version was used, but it is obsolete now as no supported Debian release supports anything other than python2.7." (From https://wiki.debian.org/Python/LibraryStyleGuide) That page also mentions "X-Python3-Version".

packaging: require hg 4.5 also for usage, not just for building

packaging: require mercurial >= 4.5 Otherwise building the doc package fails with: (third party extension evolve requires version 4.5 or newer of Mercurial; disabling) Exception occurred: File "/usr/lib/python2.7/dist-packages/mercurial/help.py", line 624, in help_ raise error.Abort(msg, hint=hint) Abort: no such help topic: evolve

evolve: clarify why returning by adding inline doc

evolve: move a code block to have right value in has_some_val Because of this we were having wrong value of has_some_val and had some buggy behavior which is being fixed by this patch. Fixed behavior is reflected by the changes in test file.

topic: add tests which demonstrate topicset failure of FilteredRepoLookupError When topics are used, it fails to find hidden revs mentioned in a revset as we are not using unfiltered view of repo. Next patch will fix it.

pick: reduce configoverride() scope Merge tool is only needed for merge.graft() (it's what graft in core hg does), so let's make the scope of the ui.configoverride() narrower.

evolve: correct action verb in a message

evolve: internationalize a message This exact message was already wrapped in _() in the same file, just not here.

evolvecmd: the proper way to deal with conflicts is to resolve them And it's worth making the suggestion an actual hint.

test-compat: merge mercurial-4.5 into mercurial-4.4

test-compat: merge mercurial-4.6 into mercurial-4.5

test-compat: merge mercurial-4.7 into mercurial-4.6

test-compat: merge mercurial-4.8 into mercurial-4.7

test-compat: merge mercurial-4.9 into mercurial-4.8

test-compat: merge stable into mercurial-4.9

branching: merge with 9.0.0

branching: merge back with stable

packaging: mark as development versions This help to distinct official release from in progress work.

Added tag 9.0.0 for changeset 756db65030c6

packaging: prepare release 9.0.0

doc: clarify the status of configuration of the obshashrange protocol

changelog: add various missing bits

templatekw: keep compatibility with Mercurial 4.5 This is simple enough to do.

test: adjust output to stable branch We are about to make a release.

branching: merge with stable

templates: don't alias keywords directly Putting the same function twice in the registry results in duplicate doc strings and therefore confusing "hg help template" output.

stack: remove 'topic.' prefix from colors/labels Stack is related to topics, sure, but it is also good enough to be recognized on its own.

stack: handle hash sizes when --debug flag is provided Since showstack() now uses fm.hexfunc() for node hashes, which depends on ui.debugflag, we should handle situations when hashes are full-sized as well.

stack: always provide (full) node hash to non-default --template

stack: remove unnecessary prefix from stack output with non-default --template "index" template keyword already exists (the current iteration of the loop), so "stack_index" it is. It follows the same convention as other template keywords, such as "files_added", "line_number", etc.

stack: provide context to formatter with non-default --template This allows using template keywords that we don't explicitly provide to the formatter, e.g. rev, branch and topic.

stack: leverage labelsgen() to produce all needed labels for fm.write()

stack: check if stack is empty more pythonically

stack: implement __bool__ and __nonzero__

stack: get stack data directly from stack and remove stackdata() stackdata function began its life in 137f8b04901e as a proto-stack: it computed stack revs on its own and only had one property to return, "changesetcount". Later it started to prepare and return more properties, but since b933a8068c17 it was computing revs using a getstack function. And then finally in 17749d9d3968 it started to rely on stack class entirely. It was a good function, but now, when all the logic it provided was factored into stack class, I'd say it's finally time for it to be put to rest.

stack: use stack._revs instead of stack.revs[1:] in external children revset The revset in question excludes all revs that are part of the stack. Using stack.revs[1:] works (rev #0 is stack base, which is not a part of the stack), but using stack._revs is technically more correct, because this variable is specifically designed to hold only revisions that are part of the stack.

test: update output to match the parent changesets This amend was forgotten.

topic: add a simple option to reject publishing The option is pretty basic but it can be used as base to build larger feature. The main target for going in this direction is to be able to distinct between user that are "simple contributors" pushing topic for review and the "maintainers" or "automation" that can publish changesets.

compat: adjust `wrapadd` for upstream Mercurial core updated the API in f385ba70e4af.

debian: override default value for SPHINXBUILD in docs/makefile This is the recommended Debian approach. As suggested by Debian Developer Stephen Kitt, use the approach recommended by Debian for building Sphinx documentation. See https://wiki.debian.org/Python/LibraryStyleGuide#Sphinx_documentation Also belated credit to Zash on the #mercurial channel on the Freenode IRC network, who suggested the Make variable approach in the first place.

obshashtree: move obshashtree in its own module The code no longer serve a core purpose. We just keep it because the command might be useful. We move it in a dedicated module so that it does not get in the way of other work.

obsdiscovery: drop `obshash` discovery protocol (issue6136) The protocol has been superseeded by stablerange for a long time. It is untested and more and more buggy. Since this is deprecated experimental code in an experimental code, we drop it. We keep the underlying computation and debug command around for now. They might still be useful to looks at repositories

cmdrewrite: use context manager for some locks and transactions These were the trivial cases where I didn't even need to extend the scope of a transaction or change indentation to fix it.

evolve: mention that --all is default, in list of options

touch: use util.acceptintervention() for closing the transaction It will close the transaction on InterventionRequired.

touch: use context manager for locks

touch: extract the logic of touching rev's to its own function This refactoring will also help us in using the context manager for locks in next patches and also help reducing the nested depth.

changelog: mention user merging in the changelog

evolve: consider using three way merge to get the user for div resolution This patch adds the logic to merge the users from two divergent csets considering the base (i.e three way merge). The case when all the three are different, we concatenate the two authors of divergent csets.

evolve: add tests for the case when div csets has different users

branching: merge with stable

evolve: move status msg to verbose mode in content-divergence resolution As suggested by Pierre-Yves, lets move these too verbose message for the user to verbose mode to decrease some noise.

debian: Override default value for SPHINXBUILD in docs/makefile Sphinx now defaults to Python 3 in Debian Buster, so we need to explicitly select the Python 2 version of sphinx-build for building docs.

docs: add SPHINXBUILD make variable This defaults to sphinx-build. The current use case is for the Evolve Debian package build.

topic: add test to increase code coverage in case of multiple topics This patch adds a test which cover the part that when we have one topic on top of other topic, this make sure that when evolving it's boundary are the current active topic (i.e hg stack)

changelog: --no-update is the default, recommend --update instead

topic: drop support for accessing csets in branch stack using bxx (issue6119) When topic extension is enabled and we have some cset whose hash is `b1234`, topic extension thinks that we are accessing 1234 cset in current branch stack. However that's not the case generally. Also I am not sure many people use this bxxx thing. Since we have a generic sxxx way to access csets, let's drop support for accessing csets using bxx which leads to bad behavior. Looking at the tests, we don't show bxxx in hg stack output anymore. I update the test to use sxxx instead of bxxxx.

legacy: drop the check for `obsolete._enabled` This looks like an ancient version check, and presumably `minimumhgversion` will do the same thing.

docs: drop references to the old temporary commit that was created with amend The sharing.rst page also mentions amend, but the diagrams aren't showing, so I'm not sure if they're in some other repo and needing to be updated.

templatekw: add `instabilities` alias for `troubles` Per https://www.mercurial-scm.org/wiki/CEDVocabulary

revset: switch internal uses and tests to `predecessors` from legacy name

revset: add `predecessor` aliases for `precursors` and `allprecursors` Per https://www.mercurial-scm.org/wiki/CEDVocabulary

revset: add `unstable` alias for `troubled` Per https://www.mercurial-scm.org/wiki/CEDVocabulary. The old name is still available (but undocumented) for compatibility. `Orphan` was renamed from `unstable` in core in 4.4, so this seems safe to do now.

docs: change `bumped` references to `phase-divergent` Per https://www.mercurial-scm.org/wiki/CEDVocabulary

docs: change `divergent` references to `content-divergent` Per https://www.mercurial-scm.org/wiki/CEDVocabulary

docs: change `troubled` references to `unstable` Per https://www.mercurial-scm.org/wiki/CEDVocabulary

docs: change `unstable` references to `orphan` Per https://www.mercurial-scm.org/wiki/CEDVocabulary

doc: correct a typo

docs: change `troubles` references to `instability` Per https://www.mercurial-scm.org/wiki/CEDVocabulary. There are a couple instances of 'troublesome' which probably need a word, as deriviatives of 'instability' didn't really fit. At least some of these are in the unlinked obs-terms.rst page.

docs: change `precursors` references to `predecessors` Per https://www.mercurial-scm.org/wiki/CEDVocabulary

docs: import figure diagrams with inkscape 0.92 No changes to the content here. This simply accepts the default option ("This file contains digital artwork for screen display") on the prompt when first opened, which is trying to upgrade from 90DPI to 96DPI.

messages: standardize on "working copy is now at" including "is" It seemed that we mostly used "is", but `hg next` and `hg prune` could print it without the "is".

evolve: move more of the progress-related variables together

evolve: delete dead calculation of number of troubled revisions The "count" variable is recalculated in the regular (no --continue) code path and it isn't used anywhere else (it used to be set to 1 in the --continue code path, but that now has its own progress counting).

evolve: drop nested locks and transactions All callers of _solveone() now take the locks and start the transactions, so there's no need to do that anymore.

next: start transaction before calling _solveone() All other callers do this, and I'm about to make _solveone() not create a transaction.

evolve: use single transaction also when continuing

evolve: run in a single transaction This speeds up `hg evolve --all` of a stack of 24 commits in the hg repo from 11.5s to 3.7s (3x). As can be seen in the tests, it also avoid some warnings about transiently unstable commits.

evolve: extract function for solving one revision, including merges Will make the next patch simpler.

evolve: mention that --all is the default behavior now

evolve: mention that not all successful operations currently update wdir

evolve: add progress support for --continue The progress starts where it left off when the previous command ran into merge conflicts.

evolve: reduce scope of progress-related variables

evolve: move progress-clearing out of _cleanup() The progress is only used with --all or --rev, so there's no need to clean it in any other case. It's therefore easier to clean it specifically in that case.

evolve: stop passing no-op "progresscb" into continueevolve() The "progresscb" doesn't do anything in the --continue case, so there's no need to pass it into continueevolve(). This lets us remove the --no-all that was added to the test as a workaround.

evolve: _solveunstable() update progress only once There's no need to draw it, then possibly write text over it (with --verbose), and then draw it again.

evolve: don't update progress just before clearing it _cleanup() will clear the progress, so there's little reason to update it just before (except for progress.debug=yes).

evolve: clean up progress bar also when using -r It looks like 1fe3da0b4601 (evolve: add --rev option to the evolve command, 2015-05-05) forgot to update the "showprogress" variable.

tests: add some basic testing of progress

evolve: increment progress *after* a whole merge commit is done The "re-stabilize" step was using the progress that was supposed to be for the next revision.

evolve: use util.acceptintervention() for closing transactions util.acceptintervention() will close the transaction on InterventionRequired.

evolve: use context manager for transactions

evolve: use consistent message and just re-raise InterventionRequired

evolve: use standard InterventionRequired instead of MergeFailure

evolve: don't use exception for local flow control The LocalMergeFailure class was added in 3f91654713dd (obsolete Move merge failure handling into stabilize code, 2012-08-20). I think the "if compat.hasconflict(r)" check was added later. Now that we have that check, we should use that for flow control instead. Note that this means that any unexpected exception from _relocatecommit() will now just raise (and roll back the transaction). I think that's an improvement.

evolve: reduce indentation with an early return statement

evolve: remove a useless return statement at end of function

test: move a test to an appropriate test file This patch introduce a new test file test-evolve-orphan-corner-cases.t for tests which covers corner cases.

test: improve the title of a test

test: remove duplicated test We have the test in test-evolve-orphan-split.t which is covering the same case this test does. Also this test states that evolve shows some error in split case, but now this is not the case we update to the top of splitted csets.

evolve: warn user for the metadata being lost in public divergence resolution This patch print warnings if some "extras" of a divergent changeset that is possibly important metadata which is being lost during the resolution of public divergence.

evolve: show a status message when merging leads to public cset I think it would be fine to tell the user that merging the two divergent csets while resolving public divergence resulted into the public cset itself. If you think we don't need to tell the user, then we can include this in debug-mode for debugging purposes at least.

test: use a shorter test name This is still long but easier to read.

stack: rename troubledcount to unstablecount

evolve: use --all by default (BC) allopt defaults to None to avoid touching _checkevolveopts(). If it were True by default, that function would always act like the flag was provided by user, and so using --abort, --stop, --continue or --rev on their own would never work (hg evolve will complain that you cannot specify both --rev and --all). A notable change in tests is "no troubled changesets" line that gets printed when using `hg evolve` to update to the evolved successor of working directory parent, but there were no actual instabilities.

tests: add --rev to hg evolve to reduce test changes for the next patch

tests: add --no-all to hg evolve to reduce test changes

evolve: don't update after instability resolution by default (BC)

tests: add --update to hg evolve to reduce test changes for the next patch

tests: don't create a nested repo for no reason

branching: merge with stable VErsion 8.5.0 got released

Added tag 8.5.1 for changeset 05c9dcf5512e

packaging: prepare version 8.5.1

test-compat: merge mercurial-4.5 into mercurial-4.4

test-compat: merge mercurial-4.6 into mercurial-4.5

test-compat: merge mercurial-4.7 into mercurial-4.6

test-compat: merge mercurial-4.8 into mercurial-4.7

test-compat: merge mercurial-4.9 into mercurial-4.8

test-compat: merge stable into mercurial-4.9

evolve: fix progress display with hg<4.7 Since d1aab9d82f5b (evolve: adapt for deprecated ui.progress(), 2019-01-14), the "item", "units" and "total" are all set to default values. The last one is especially bad since it results in a "spinner" instead of a progress bar.

packaging: mark extensions as tested with 5.0

branching: merge with stable

test-compat: merge mercurial-4.5 into mercurial-4.4

test-compat: merge mercurial-4.6 into mercurial-4.5

test-compat: merge mercurial-4.7 into mercurial-4.6

test-compat: merge mercurial-4.8 into mercurial-4.7

evolve: use merge.update() for resolving phase divergence Iterating over the manifest when tree manifests and narrowness is in play produces entries for paths outside the narrowspec that represent trees. For example, if the tests/ directory of the hg repo was not in the narrowspec (and the hg repo was using tree manifests, which it doesn't), then there would be a "tests/" entry in the manifest. The merge code deals with some of these cases. For example, it's valid to do a merge if only the local side changes directories outside the narrowspec. That allows rebasing a local commit onto a public commit that had changes to the excluded paths to work. However, _resolvephasedivergent() was iterating of the manifests, which resulted in crashes for some of our users when they tried to resolve phase-divergent commits (actually content-divergent commits that became phase-divergent after the intermediate rebase). We can fix that by relying on merge.update(), since that already handles this case. [This re-install a changeset backed out for compatibility reason]

test-compat: merge mercurial-4.9 into mercurial-4.8

compat: drop support for Mercurial 4.4

test-compat: merge stable into mercurial-4.9

compat: backed out changeset cfcb7eedc666 That changesets breaks 4.4 compat. We will reinstall it for the next version.

test-compat: adapt test output to 4.4

test-compat: merge mercurial-4.5 into mercurial-4.4

test-compat: merge mercurial-4.6 into mercurial-4.5

test-compat: merge mercurial-4.7 into mercurial-4.6

test-compat: merge mercurial-4.8 into mercurial-4.7

test-compat: merge mercurial-4.9 into mercurial-4.8

test-compat: merge stable into mercurial-4.9

compat: fix a typo in compat patch

changelog: mention extensions isolation (issue6121)

topic: only wrap wireprotocaps for repo with topic This helps repository with and without topic to coexist in the same process.

topic: only wrap _pushb2phases for repo with topic This helps repository with and without topic to coexist in the same process.

topic: only wrap handlecheckheads for repo with topic This helps repository with and without topic to coexist in the same process.

topic: only wrap wireprotobranchmap for repo with topic This helps repository with and without topic to coexist in the same process.

topic: only wrap _headssummary for repo with topic This helps repository with and without topic to coexist in the same process.

topic: only wrap desthistedit for repo with topic This helps repository with and without topic to coexist in the same process.

topic: only wrap _destupdatetopic for repo with topic This helps repository with and without topic to coexist in the same process.

topic: only wrap _destmergebranch for repo with topic This helps repository with and without topic to coexist in the same process.

topic: only wrap gettopicrepo for repo with topic This helps repository with and without topic to coexist in the same process.

topic: only wrap mergeupdate for repo with topic This helps repository with and without topic to coexist in the same process.

topic: only wrap committext for repo with topic This helps repository with and without topic to coexist in the same process.

topic: only affect `hg commit` behavior for repo with topic This helps repository with and without topic to coexist in the same process.

topic: only wrap workingctx.__init__ for repo with topic This helps repository with and without topic to coexist in the same process.

topic: introduce a `hastopicext(repo)` function This function check if a repository has the extension enabled or not.

tests: consolidate the public-content-divergent test cases This commit move the public-content-divergent tests in dedicated test files.

evolve: improve inline documentation for the public divergence resolution

cleanup: remove a duplicated line

changelog: mention topic compatibility

branching: merge stable back into default

tescompat: revert test change for mercurial-4.9 We should still test against mercurial-4.9 expected output.

branching: merge 8.5.0 expected output in stable Upstream stable is now for mercurial 5.0

testcompat: open a branch dedicated to mercurial 4.9 Upstream merged its default branch into the stable one to prepare for Mercurial 5.0

topic: add compatibility for writing copy metadata in changelog.add() As seen in hg 0e41f40b01cc.

serverminitopic: add compatibility for branchcache that now uses self._entries A copy of 9b0cde5efbc9.

branching: merge stable into default

packaging: mark version as development version Thsi help to reduce confusion for people using evolve from source

Added tag 8.5.0 for changeset 33c617626fd9

packaging: prepare version 8.5.0

test-compat: merge mercurial-4.5 into mercurial-4.4

branching: merge stable back into default

test-compat: merge mercurial-4.6 into mercurial-4.5

test-compat: merge mercurial-4.7 into mercurial-4.6

test-compat: merge mercurial-4.8 into mercurial-4.7

test-compat: merge stable into mercurial-4.8

changelog: merge 8.4.1 and 8.5.1 entries

branching: merge stable into default

branching: merge default change into stable

branching: merge default branch in stable to prepare 8.5.0 release The release is not fully ready, but the merging is non-trivial.

uncommit: abort if an explicitly given file cannot be uncommitted (BC) This corresponds to f4147ca63d39 in the core uncommit extension.

evolve: store --update flag state in evolvestate file Part of 9c522bdfc846 that wasn't yet accepted.

state: implement cmdstate.__contains__()

changelog: mention Martin work on phase-divergence

test: revert update from 5.0 record logic This revert the test changes from 90f94231db5d.

tests: remove note about already implemented --confirm with --continue 4544067b831b made --confirm work with --continue, but didn't update/remove this comment.

evolve: use merge.update() for resolving phase divergence Iterating over the manifest when tree manifests and narrowness is in play produces entries for paths outside the narrowspec that represent trees. For example, if the tests/ directory of the hg repo was not in the narrowspec (and the hg repo was using tree manifests, which it doesn't), then there would be a "tests/" entry in the manifest. The merge code deals with some of these cases. For example, it's valid to do a merge if only the local side changes directories outside the narrowspec. That allows rebasing a local commit onto a public commit that had changes to the excluded paths to work. However, _resolvephasedivergent() was iterating of the manifests, which resulted in crashes for some of our users when they tried to resolve phase-divergent commits (actually content-divergent commits that became phase-divergent after the intermediate rebase). We can fix that by relying on merge.update(), since that already handles this case.

evolve: compat patch for recordfilter change in mercurial This patch fix the broken things because of upstream changes in recordfilter() which is being used to select the hunks interactively. It fixes the test-uncommit-interactive.t by adding the compat layer. But for test-split.t I had to fix the tests manually. To make it more clear: splitting broke at evolve side because after that upstream change now interactive mode doesn't prompt "examine change to foo" if foo is mentioned explicitly using cli; and directly jumps to hunks selection prompt (well, only if there is any changes at hunks level) And the main issue is when file which is explicitly mentioned has no changes at hunk level (for e.g copy, rename, mode change, empty new file), because in that case you don't have any control on selection of that file and it would be included automatically in first cycle of interactive selection. And this "no changes at hunks level" was the reason for test-split.t breakage as now it didn't prompt for those files which are passed on cli. To fix this I have included some content in those files to make sure that tests still demonstrate the same behaviour as they were doing before breakage. Also, I replaced some "n" with "s" as it make more sense to skip all the changes to that file in one go instead of hitting "n" multiple times (if there were multiple hunks)

test: add file cancelation scenario for phase-divergence This improve the test coverage for corner cases.

test: introduce a more complete test for phase-divergence The test the preservation of "change" cancellation in multiple situation.

test: rename test about content divergent meta data The test now use a name in line with the other content divergence test cases.

test: cleanup title and doc phase-divergent cases

test: remove remaining duplicated file This file should have been removed in d02a7c8cdee5.

test: add test for content-divergence with file removal File addition/removal tends to introduce corner case. The actual test content comes from a Martin von Zweigbergk changesets.

test: clarify changeset concent in the first content-divergence test

branching: merge with stable merge with new changeset on stable.

branching: merge with test consolidation Conflict were simple once isolated.

branching: merge with stable This merge happens before the large test consolidation on stable.

compat: drop getcachevfs, repo.cachevfs is supported in hg 4.4

evolve: when using evolve, cache files belong into the master repo

changelog: mention issue6089 fix

pick: update working dir branch (issue6089) Previously the working copy was left in an "inconsistent" state.

tests: rename test-grab to test-pick This match the new command name.

evolve: make sure div resolution don't replace the initial author (issue6113) Before this patch, divergence resolution logic taking the current user who is running the resolution command as the author of commit that will be created after merging the two divergent csets. This patch make sure that we preserve the initial author. Changes in test file reflect the fixed behaviour.

evolve: add test which demonstrate issue6113 This patch introducing a new test file which contains the test for issue6113 which is `test-evolve-content-divergent-distributed.t`. This test file is supposed to contain tests for distributed workflows. Next patch will fix the bug.

test: consolidate the content-divergent test cases This commit move most of the content divergent related test in dedicated and thematic test files.

changelog: mention the fix to issue6057

extensions: avoid enabling evolve for unrelated repositories (issue6057) The support for the `experimental.evolve` config have been there forever. So we can stop using the global variable. This help avoiding evolve to enable itself of repository that did not explicitly activated it. For example, this is useful for hgwebdir. Before this change, evolution feature had to be explicitly disabled or enabled in every single repository.

stack: improve set combination logic

stack: fix phasecache._phasesets check logic When _phasesets is not None, it's a list, and it contains set()s of revisions in a specific phase, starting from public, draft, secret and so on. But since repos are supposed to have the majority of revisions in public phase, the first element of this list is not a (potentially huge) set, but None. Previously this code tried to check if there's any element that is None, and was always finding None at index 0, so the short path was executed every time and the rest of the function was never used.

evolve: remove todo markers from a test which has been fixed

next: clarify tests related to `commands.update.check` test This clarify the test and help understand the next changeset.

stack: simplify phasecache checks in _stackcandidates()

evolve: fix divergence resolution when not merging a descendant In divergence resolution, when we merge other cset with divergent one we pass a `mergeancestor` arg to merge.update() and before this patch we were passing `mergeancestor` as True in every case. But it should be True only when we are merging a descendant onto an ancestor. When mergeancestor is True it does two things: 1) allows the merge if the destination is the same as the parent of the ctx (so we can use graft to copy commits) 2) informs update that the incoming changes are newer than the destination so it doesn't prompt about "remote changed foo which local deleted". So this patch change it to pass `mergeancestor` as True only when it is required. And changes in test file shows that it wasn't prompting either in those cases when it should (acc. to 2nd point) Test written by Pierre-Yves David, based the one updated in 5dbaabfe2c59.

test: fix a phase divergence test case to actually not contains conflict The previous test case had a conflict, but that conflict is currently not raise by phase divergence resolution. This is an independent bug to solve.

tests: use current instability names everywhere

test: apply unstability resolution to `test-unstability-resolution-result.t` The test cases in this file probably deserved to be dispatched in other more thematic test file.

test: rename and clarify content of test-unstable-content-divergent.t The troubles → instabilities renaming was not applied here.

test: rename and clarify content of test-unstable-orphan.t The troubles → instabilities renaming was not applied here.

evolve: refactor the code which swap two nodes (in divergence resolution) I see that we are using the code to swap the "divergent" and "other" node; and updating the evolvestate accordingly at many places in divergence resolution. This patch extract that code to a function. Although I don't have strong opinion on this patch. It's a volunteered patch to remove some redundancy.

evolve: remove tmpctx arg from _resolvephasedivergent() Before this patch, in _resolvephasedivergent() function we were passing an optional argument "tmpctx" which was used to store a rebased version of bumped if required. But for sake of simplicity we can achieve the same results even without passing this "tmpctx" into that function. No changes in test files demonstrate that this patch don't break anything.

doc: clarify _resolvephasedivergent role and constraint Small documentation patch before larger changes.

branching: merge with stable

evolve: unindent some lines caught by flake8 The message was: "E117 over-indented".

obsexchange: avoid importing io/StringIO module with a different name Since we're not interested in the whole module, let's shorten the imports and aviod the lint warnings. Otherwise pep8-naming complains: - import io as StringIO: N812 lowercase imported as non lowercase - import StringIO as io: N813 camelcase imported as lowercase

evolve: remove unused StringIO import

compat: add support for new arg name in memfilectx.__init__ This makes us compatible with 550a172a603b (memctx: rename constructor argument "copied" to "copysource" (API), 2019-03-19).

topic: make topics compatible with branchmap refactor See https://phab.mercurial-scm.org/D5290

evolve: compat patch to fix broken evolve obsdiscovery This patch preserve the compatibility for the changes upstreamed by e5ece0f46b40 in hg-committed.

tests: add obslog output to some tests for divergence resolution As suggested by Boris Feld.

tests: add tests for divergence resolution with split+amend As suggested by Boris Feld. G: changed tests/test-evolve-phase-divergence.t

test: use 'split' instead of 'splitted' in the tests

evolve: drop the "computing new diff message" for phase-divergence Future code update will remove this message, dropping it early clarify the future diff.

evolve: clarify output when phase-divergence resolution results in no change I found it pretty confusing that it said "committed as abcd1234" when no commit was created.

tests: reduce dependence on previous tests for setup The tests in test-evolve-phase-divergence.t built on each other for no good reason, which makes the tests hard to modify and hard to read. This patch cleans up some of them. We should clean up the rest too, but the rest won't matter for my next patch, so I haven't bothered.

branching: merge with stable

topicmap: add compatibility for branchcache that now uses self._entries

evolve: fix over-idented block

tests: create "private" repository in already-updated state I'm going to add more copies of the repo and it's easier to work with the repo if it's already been updated.

branching: merge with stable

touch: use action "touch", not "<function touch at ...>", in precheck I just got this message: abort: cannot <function touch at 0x7fde174b69b0> public changesets: f177c51ae98f It has been this way since 1b7b9acda2a9 (touch: use precheck to validate revision, 2017-07-23).

prune: use action "prune", not "touch", in precheck We have been using "touch" since 06ee4ec88190 (prune: use precheck to validate revision, 2017-07-23). That commit modified the tests, so it kind seemed intentional, but I suspect it was still just an oversight.

evolve: test a common case of public divergence evolve: in pubdiv, handle the case when public branch is behind the mutable In public divergence, when public branch is behind the mutable one the behaviour that this patch added is that we relocate the draft one to public side. This should be fine in most case, but might be an issue in some other (eg, when the draft side of the divergence is rebase away from ancestors it relies on.) Test written by Pierre-Yves David <pierre-yves.david@octobus.net>

evolve: add condition in case of pubdiv to not swap the two div csets In public divergence, we have pinned the public cset to the local side when merge is performed. And in divergence resolution implementation divergent cset is the one who is kept at the local side, so to say this in one line: we have pinned public cset to the divergent side. However, in some cases we need to swap the "divergent" and "other" but we can't swap them in case of public divergence as thought would cause an error because of not being able to relocate public cset.

evolve: add logic to handle a particular case in divergence resolution Case is: In content divergence, to check if the two csets are on different parent or same parent what we do is find gca of the two csets and then compare it with the singlesuccessors of two csets's parents. Now, if gca is same as singlesuccessors of both the csets then we conclude that both are on the same parent which is not totally correct as it is possible that both the parents of two csets are obsolete with a single successor. This patch adds the logic to cover the above metioned case correctly, deciding the part when we need to relocate a cset to other side while resolving the stack of divergent csets. Changes in test file demonstrate that before this patch we were not doing it right because it should have relocated every node to other side instead of just relocating the first one from the stack.

evolve: add a test reflecting wrong behaviour of content-div resolution This patch shows that currently the case when stack of divergent csets have different parents is not handled correctly. For a stack of csets which has different parent, right behaviour is to relocate every node before performing merge but present implementation only relocate first cset in the stack and skip relocation for the rest of csets, this is not what we expect it to do. Next patch will be about fixing this issue.

compat: add support for new arg name in memfilectx.__init__ This makes us compatible with 550a172a603b (memctx: rename constructor argument "copied" to "copysource" (API), 2019-03-19).

evolve: preserve compatibility for hg < 4.8 versions This patch make sure evolve don't break if user using a version of hg before this 7694b685bb10 patch on hg-committed.

evolve: warn user if cset desc is being lost In public divergence when merging results in same as public cset, we only create markers to solve the divergence i.e. [(public, (other,))] In this case if other cset had a desc which was different from public that will be lost. So this print out a warning message to user that desc of non-public cset is being lost. Changes in test files reflect the added behaviour.

changelog: mention latest improvement

stack: make @ (current) more important than $ (some sort of unstable) Base is still ^ even if it's currently checked out.

stack: show content and phase divergent state and symbol Replacing ' ' with '-' is just for color labels.

stack: handle external-children just like other states This also fixes color labels: previously when external-children was combined with another state, --color=debug would show something like: [topic.stack.index topic.stack.index.current - external-children|s4] Now there is a separate label for each state.

stack: optimize revset used for stack --children Since that option needs to exclude changesets on the current stack, let's just use stack.revs (that are cached) for that and skip branch() or topic() revset. st.revs[0] is the stack base, we don't consider it actual part of stack.

stack: rename unstable -> orphan That's the only instability stack can currently show, but it's about to show more and showing "unstable" for one type of instability just isn't good.

evolvecmd: compatibility with older geteffectflag() that took a tuple Before bae6f1418a95 the function took a tuple of 2 arguments instead of these 2 arguments directly.

obshistory: remove unused geteffectflag(), it's been in core since 4.4

compat: remove old vocabulary change fallbacks Mercurial 4.4 (minimal hg version evolve currently supports) already uses the new names and in fact has deprecation warnings saying that 4.4 is the last version of Mercurial to support the old names.

branching: merge with stable

evolve: fix a typo from directory to dirty

evolve: fix the dirstate after setting parents with dirty wdir (issue5930) Before this patch, during an interrupted evolve it can be confused about copies. This patch fix the dirstate when parents are changed using dirstate.setparents() with a dirty wdir. Changes in test file reflect the fixed behaviour.

evolve: add tests which shows evolve can get confused about copies (issue5930) This test demonstrate that an interrupted evolve can get confused about copies. Next patch will fix this.

test: include some documentation to describe the case briefly

split: use ui.configoverride to preserve phase while commiting As we need to preserve the phase of revision which is going to be splitted into multiple revision. This patch use ui.configoverrides to temporarily overwrite the phases.new-commit config option at the time of committing a new node.

branching: merge with stable

changelog: add entries for the recent fixes

evolve: make sure user can recover from conflict (issue6053) we make sure while continuing interrupted evolve it reach to right code This patch fix the issue6053 which is about falling evolve into unrecoverable state and this is how it call fall into it: If after some rewrite operations we reach to a point where resparent for two content divergent csets is obsolete and we hit a conflict in relocation, so now if you run `hg evolve --continue` after resolving the conflict: code execution would check if wdir parent is obsolete and run solveobswdp() rather than going to continueevolve() part and there we will hit by a "abort: outstanding uncommitted merge" And same for `--abort` or `--stop` flags. Changes in tests reflect the fixed behaviour.

evolve: add test which shows unrecoverable evolve state (issue6053)

evolve: add a test in pubdiv where only merging leads to conflict Added a test in public divergence where both the cset are on the different parent and no conflict in relocation but in merging.

evolve: cover continue case in pubdiv when merging results in same as public This patch is the part of a series which is adding support of public divergence resolution of the case when merging results in same as pubic cset. Changes made in this patch cover the --continue flag for the case: when we had conflicts in merging. Modified lines in previously added tests reflects the changed behaviour. Also added some additional tests. I guess I have covered every possbile case in test-evolve-public-content-divergent.t While reviewing, if you see any test missed please let me know. (I think there is one test missed that I will send as a follow-up)

evolve: handle a case in pubic-div when merging results in same as public In public divergence resolution, what we do is: 1) first apply content divergence resolution 2) then phase divergent resolution on resultant node of 1) case While doing case 1 it is possible that result of merging the two csets would have same changes as public one contains. And then processing the case 2 would create an empty commit which is not something we want to do. So this patch catch that same case when merging results in same as public cset and don't create a new node, instead to solve the divergence it just add a obsmarker from "other divergent" to "public divergent" i.e. [other, (public,)] Next patch will add the continue case handling for this same case. This patch also adds the tests for the different cases which are possible for the above mentioned case. There is test for continue case too which is broken in this patch and will be fixed in next patch.

evolve: rename npublicdiv to publicnode In next patches it would make more sense to use publicnode instead of npublicdiv.

evolve: in pubdiv resolution make sure that transaction get closed Before this patch it could be possible that transaction won't get closed if res is False in the pubic divergence resolution when continuing the hg evolve.

evolve: pin the public cset to local side in merging when solving public div While working on public divergence, I think it would be better to pin the public cset to local side of merge for understanding and handling the future cases without any ambiguity. Changes in tests reflect the changed behaviour.

evolve: add tests of relocation case in public divergence These tests cover the cases when "other" is behind the "public" one and we would have to relocate "other" here. It covers all the conflicting and non-conflicting cases.

branching: merge with stable

split: make sure hg split preserve the phase of splitting cset (issue6048) Before this patch, while splitting we were not taking the phase of splitting cset into account. By default new commits would have draft changes. As we know there can two possible phases i.e draft and secret in rewriting csets thing, so this adds the handling of secret phase (as default is draft) Changes in test file reflect the added behaviour.

split: add tests which demonstrate the issue6048 Next patch will fix this issue.

evolve: fix the broken behaviour on div resolution in relocation case In content divergence resolution, I see that when "relocationreq" is True which means both the cset are on different parent and relocation is required. In this case when "divergent" is the one who is behind the "other" cset in DAG, we swap them. At this point one thing we missed is to update the evolvestate['divergent']. Because of this in continue case we didn't get the right value of obsmarkers creation. In this patch I added the code to update the evolvestate when we swap them. Now as we know "divergent" and "other" can be swapped in some cases, it is better to store the intial divergent separately in evolvestate which is evolvestate['orig-divergent'] and later this value is used to update the evolvestate['replacements'] which is the track of revisions which has been resolved. Changes in tests demonstrate the fixed behaviour.

evolve: change in a test which demonstrate divergence resolution can break This patch just add "--any" flag to "hg evolve --content-divergent" in a test and changed output reflect that there is something which is not handled correctly. After looking into code I found that it is because of using a evolvestate value which is not updated i.e evolvestate['divergent'] Next patch will fix this problem.

tests: update the title of a test as it's not what it says I see that this test has conflict in relocation and in merging also. We do have a test for the case when only relocation leads to conflict.

obsexchange: avoid importing io/StringIO module with a different name Since we're not interested in the whole module, let's shorten the imports and aviod the lint warnings. Otherwise pep8-naming complains: - import io as StringIO: N812 lowercase imported as non lowercase - import StringIO as io: N813 camelcase imported as lowercase

evolve: remove unused StringIO import

evolve: unindent some lines caught by flake8 The message was: "E117 over-indented".

branching: overwrite broken merge with a correct one Thanks goes to Anton Shestakov for spotting and fixing this.

branching: merge with stable

branching: merge with stable

obshistory: omit keyword arguments with default values in patch.diffui() calls d4c9eebdd72d made that function not take "relroot" and "prefix" anymore, and instead expect "pathfn". These two chunks of code in this patch simply generate patches for obslog -p and they broke because they tried to provide default values for "prefix" and "relroot" for no apparent reason (and also for "changes"). Let's instead rely on defaults of patch.diffui().

branching: merge with stable

test: wider another time matching in prev/next lock testing There does not seems to be a reasons for specifically matching 4-6 seconds. Wider matching will avoid intermittent failure.

branching: merge with stable

test: wider time matching in prev/next lock testing There does not seems to be a reasons for specifically matching 4-6 seconds. Wider matching will avoid intermittent failure.

topic: make topics compatible with branchmap refactor See https://phab.mercurial-scm.org/D5290

prompts: use 1-indexing in revselectionprompt() The consumer is generally a human, not a computer. I've made the mistake of entering "1" instead of "0" a few times myself.

test-split: split now ignores diff.* settings From 66399f2e92aa: "It is assumed by the author that the `[diff]` section is primarily for *viewing* diffs, and that it is unlikely what people intend when attempting to commit or revert."

pick: add --tool for hg pick to specify mergetool This patch used configuration override to enable you to pass a mergetool via the --tool flag in case there is any merge conflict.

revset: use getintrange() to parse relation subscript After 13f7a6a4f0db parsing the tokens into usable values needs to happen inside subscript relation functions.

branching: merge with stable

compat: pass contexts to _findlimit() (issue6066) Now _findlimit() in copies module accept context instead of rev number. See https://phab.mercurial-scm.org/D5595

packaging: fix debian changelog

changelog: add relevant entries

evolve: add test for the case where public divergence on difference parents This test make sure that public divergence resolution also handle the case when divergentes changeset does not share the same parent at one end because one end simply rebase. The other side has actual an actual diff change. For now this would work only for the case when we need to relocate the mutable one. Other case is still left to work on.

evolve: update the public divergence resolution logic to cover --continue case To continue the interrupted evolve in case of public divergence: we will have to store the node of that public cset which was in content divergence with the other cset, so that we can perform the phase divergence resolution between that public cset and a newnode which is the result of content-divergence resolution. Added tests reflect the behaviour this patch adds.

evolve: add logic to resolve content-divergence with a public cset Public content-divergence is the divergence where a cset is content-divergent with a public cset. To solve public divergence: 1) perform content-divergent resolution 2) resultant node is phase divergent 3) perform phase divergence resolution It is the intial logic to solve public divergence. Next patches will be the covering the cases of: 1) relocation: when we need to relocate to node 2) continue: to add this resolution in case of --continue 3) the case when content-divergence resolution gives a result similar to public cset. Added test-evolve-public-content-divergent.t reflect the added behaviour.

evolve: introduce function to create a obsmarker relation even for public cset To create obsmarkers we use obsolete.createmarkers() function, but because of security reasons this function refuses to create obsmarkers for public cset. And we need to create obsmarkers for a public cset while solving public content divergence. So introducing this function which create a obsmarker relation even for immutable cset. Currently this function create obsmarker for a single relation, in contrast of obsolete.createmarkers() which create markers for multiple relations. Upcoming pathces will be using this function.

topic: make ranges work in revset relations like 'foo#stack[1:2]'

topic: simplify #stack index check/access Also using stack.revs instead of list(stack). It's equivalent and stack.revs is @propertycache'd.

topic: make ranges work in revset relations like 'foo#topic[1:2]' Since #topic is very similar to #generations, we reuse the function directly. Few tests because #generations is already tested in core.

evolve: add description of function solveobswdp

evolve: document the "if not shouldupdate" block I missed the "not" in "if not shouldupdate: ... hg.updaterepo()" and it took a long time for me to realize it's about updating *back* to the original commit (or to its successor). Let's add a comment to maybe help the person not make the same mistake.

evolve: make "startnode" consistently be a node, not a context This fixes a bug where we didn't print "working directory is not at ..." after `hg evolve --continue` because we were comparing a node and a context.

evolve: use context managers for locks and transaction

prune: use scmutil.cleanupnodes()

uncommit: use scmtutil.cleanupnodes()

amend: use scmutil.cleanupnodes() with --patch

pick: use scmutil.cleanupnodes() A side-effect is that `pick` should now also move bookmarks.

fold: use scmutil.cleanupnodes() scmutil.cleanupnodes() has existed since 4.3 and had the metadata argument since 4.4, so we can now safely depend on it.

revset: subscriptrelations functions now have two bounds

branching: merge with stable A new version has been released

packaging: mark as development version

Added tag 8.4.0 for changeset 8c4289d0e91e

packaging: prepare release 8.4.0

minitopic: add warning about extensions state and testing

test-compat: merge mercurial-4.5 into mercurial-4.4

test-compat: merge mercurial-4.6 into mercurial-4.5

branching: closing test-compat branch for mercurial-4.3 We are dropping support for this version.

test-compat: merge mercurial-4.7 into mercurial-4.6

test-compat: merge mercurial-4.8 into mercurial-4.7

test-compat: merge stable into mercurial-4.8

changelog: explicit the part related to topic

branching: merge with future version

branching: start branch for test compat with Mercurial 4.8

test-compat: reverse test change from 9bce7e6c18b3 This change are 4.9 only

test-compat: reverse test change from db87432df8de This change are 4.9 only

test-compat: reverse test change from ac39cfb5bddd This change are 4.9 only

branching: merge stable into default prior to new version

branching: merge wit stable

obsdiscovery: use consistent filtering during discovery If we computed revision using an unfiltered repository, we should run the discovery unfiltered too.

evolve: make sure we don't stop after evolution of obsparent (issue5881) Before this patch, we were not handling the case when revset given by user also contain wdir parent which is obsolete. We were explicitly checking if wdir parent is obsolete, then update to its successor and return; without checking if we still have some revisions to be evolved. This patch make sure that after it update to the successor of obsolete parent it doesn't stop and continue to evolve other remaining revisions. Changes in tests/test-evolve-issue5881.t reflect the fixed behaviour.

evolve: extract "update" code into its own function This patch moves the code to a function which handles the case when wdir parent is obsolete and evolve will just update to its successor. And extraction of this to a function will also help us to fix the issue5881.

evolve: add test case for issue 5881

evolve: add test which shows fixed behaviour of `hg evolve` (issue5686)

evolve: cover the orphanmerge part missed by lastsolved Let's cover the orphanmerge part also as it was missed at the time when we introduced "lastsolved". Chagnes in test-evolve.t reflect the fixed behaviour.

evolve: add test which shows orphan-merge resolution can result to crash It happens when there are still some revisions to evolve after evoluiton of a orphan-merge. And the reason behind crash is that `lastsolved` is not being updated and contain a wrong value which is now hidden. Next patch will fix this problem.

next: populate stacktemplate to the missed _nextevolve function

evolve: adapt for deprecated ui.progress() This makes a wrapper that has an interface like ui.progress() even though that's the old interface that will eventually go away. That's because it's a lot easier to convert from the ui.progress() to the ui.makeprogress() interface than the other way around. Once we no longer support hg<4.7, we can migrate our callers to ui.makeprogress() and drop the wrapper.

fold: sort concatenated commit messages in topological order It seems better to have the proposed commit message look the same whether the user ran `hg fold --exact -r . -r .^` or `hg fold --exact -r .^ -r .`.

topic: make revsets like 'foo#topic[0]' work Essentially, 'topic' relation subscript function is the same as 'generations', but is limited to revisions marked with one particular topic.

changelog: add an entry about the revset stack indexing

test: adapt output to minor behavior change CORE-TEST-OUTPUT-UPDATE: bad05a6afdc8

tests: remove unused variable and needless subshell in testlib/pythonpath.sh

branching: merge with stable

tests: conditionalize some output for Windows A couple of these were already present, but out of place.

test-pyflakes: use '/' when printing the files to check If paths are printed with '\', `xargs` sees the parts allsmooshedtogether, and pyflakes errors out.

tests: fix PYTHONPATH manipulation on Windows Without the semicolon separator and quotes, the variables ends up like this: c:\Users\Matt\projects\hg-evolve:c:\Users\Matt\projects\hg-evolve: c:\Users\Matt\projects\hg;c:\Users\Matt\projects\hg-evolve\tests; c:\Users\Matt\projects\hg\tests That in turn makes the local evolve code unreachable, and the system installed code is tested instead. I'm testing against `uname` instead of a trivial python script printing os.name because maybe one day tests will run on WSL.

exthelper: update documentation with some examples for using registrar aliases

exthelper: simplify the ability to register filesets Same mechanism as revsets earlier in this series.

exthelper: simplify the ability to register templates Same mechanism as revsets earlier in this series.

exthelper: simplify the ability to register revsets I think this is what Yuya and Boris agreed on.[1] This happens *after* the extsetup phase now (and after the _aftercallback notifications). But this is trivial, mergeable between exthelper instances, and doesn't need to have the extension name supplied when registering. [1] https://www.mercurial-scm.org/pipermail/mercurial-devel/2018-December/125888.html

exthelper: support the option argument when registering a command Largefiles uses this 5th argument with push and pull, so this will be tested in the next commit. I assume the reason for unrolling and reforming the tuple in each finalxxxsetup() is to validate that something proper was passed in when registering. But it's better to explode when decorating than during the delayed actual registration.

exthelper: simplify configitem registration

exthelper: avoid a [] default arg to wrapcommand()

configitems: move the default values to the registrar decorator The exthelper refactoring of configitems stopped registering dynamicdefault for everything, which in turn causes a lot of warnings about mismatched defaults. There are a few stragglers- I didn't fix topic-mode because I wasn't sure what the default should be, and the alias for odiff was explicitly given a default for compatibility.

topic: make revsets like foo#stack[-1] work too We remove base commit when n < 0 because it's confusing to see base (which technically isn't a commit inside the stack) at both 0 and -n.

topic: make revsets like 'foo#stack[0]' work 'stack' relation subscript function is another way to refer to s0, s1, etc. But instead of aborting in many cases it will simply return an empty set.

exthelper: remove '_' in final_xxx() function names to conform to hg style

branching: merge with stable

packaging: mark as development version This avoid confusion when installing from source.

Added tag 8.3.3 for changeset 97b18934d6db

packaging: prepare version 8.3.3

test-compat: merge mercurial-4.4 into mercurial-4.3

test-compat: merge mercurial-4.5 into mercurial-4.4

test-compat: merge mercurial-4.6 into mercurial-4.5

test-compat: merge mercurial-4.7 into mercurial-4.6

test-compat: merge stable into mercurial-4.7

evolve: detect unresolved conflict during evolve --continue (issue5966) Evolve --continue was ignoring unresolved conflict, proceeding with its operation without any check. This is now fixed. This test case was built and formalized by Matt Harbison and Anton Shestakov.

tests: filter out uninteresting log events Core has new ui.log() calls added (for tracking extension loading and performance, for example), but in test-discovery-obshashrange.t it produces a bunch of lines that are unrelated to the test. So we filter them out to not touch test output. This is inspired by 1de3364320af in core.

test-compat: merge mercurial-4.4 into mercurial-4.3

test-compat: merge mercurial-4.5 into mercurial-4.4

test-compat: merge mercurial-4.6 into mercurial-4.5

test-compat: merge mercurial-4.7 into mercurial-4.6

test-compat: merge stable into mercurial-4.7

linter: silence rightful complains about unused variable

topic: merge _showlasttouched logic into _listtopics Makes --age work with --verbose.

topic: simplify _showlasttouched and _getlasttouched (and its data structure)

branching: merge with stable

split: add a --interactive flag While the interactive mode is useful for end-users, there are some cases where they don't need the interactivity overhead. Moreover, this allow the split command to be used in automated scripts.

split: take file patterns to limit selection on matching file patterns When splitting a big changeset, the user often wants to extract specific files from the changeset while leaving the others files intact. The split command can now be called with multiple file patterns. Only files that match one of the patterns will be shown in the interactive selection. The user will be able to create one or more commits with those files. When the user finishes splitting the matching files, a last commit will be made with the remaining of the files.

split: only accept explicit revision through the `-r` option We are about to introduce new non-optionnal arguments to defines files patterns. To avoid confusion and to be closer to the standard options, only accepts revisions with the `-r` option.

issue-6028: resolves 6028 - return (False, ".") instead of (False, '') Since 4.8, repo[] no longer takes an empty string, so we update to a valid with similar semantic. This value is strange and we should probably use an actual node ID here, however, this is an independent issue.

branching: merge with stable

tests: add sparserevlog cap since hg now has sparse revlogs by default CORE-TEST-OUTPUT-UPDATE: 3764330f76a6

split: rework the prompt to be more standard We dropped the initial capital and moved to a positive question.

split: add a option to discard remaining change during split This is useful to drop unwanted remaining change spotted during the split.

split: add a way to abort a split from the prompt This prompt is asking the user about the next step. Offering to abort there make sense.

split: add a help entry to the final prompt We are about to add more option in this prompt. Adding a way to document then seems like a good first step.

topic: make revset argument messages be similar to such messages in core

tests: don't forget to add hg serve pid to DAEMON_PIDS in test-wireproto.t Otherwise it leaves a process running after each run-tests.py launch.

test: adapt to change output change on default We now have a `wcache` directory in `.hg`.

safeguard: don't reimplement auto-publish if it's in core (will be in 4.9)

safeguard: the new default for auto-publish is 'publish'

stack: add a --children flag to stack If a commit in the stack has been merged outside the current topic, show a trailing status of (external-children)

extensions: avoid shadowing imported module names with exthelper methods The previous code worked, but confused pyscripter.

extensions: add fileset support to exthelper This will be tested by using it in lfs.

extensions: add uipopulate() support to exthelper

exthelper: clarify copyright and licence for the module The exthelper is quite useful and Matt Harbison is looking into having it upstreamed. We clarify the legal situation first.

packaging: merge with version 8.3.2

packaging: mark development version This help reduce the confusion when people install from source.

Added tag 8.3.2 for changeset 7edc5c148df0

packagin: prepare release 8.3.2

test-compat: merge mercurial-4.4 into mercurial-4.3

test-compat: merge mercurial-4.5 into mercurial-4.4

test-compat: merge mercurial-4.6 into mercurial-4.5

test-compat: merge mercurial-4.7 into mercurial-4.6

test-compat: merge stable into mercurial-4.7

changelog: add information about the topic message improvements

branching: merge with stable

evolve: update the abort to a ProgrammingError This error should be a ProgrammingError and as Pierre-Yves David said ProgrammingError type did not existed when this code was written, so lets update this now.

topics: improve the message around topic changing The new message says 'cleared' when topic name is cleared, if we change to a new topic name, it mentions the new topic name too.

tests: fix for POSIX compliance * Mac ls will exit with a different message and exit code [1], replace with grep (which is what mercurial tests do) * POSIX echo has no -n switch, use printf instead

style: remove unused variables Both `as e` references are not used and can go.

evolve: fix a documentation typo

compat: drop 4.3 related obsfate code All the necessary piece are supported starting 4.4.

compat: drop code dealing with incompatibility for --publish with 4.3 We dropped 4.3 support.

compat: clarify when some conditional are used The middle condition was an intermediate state that never existed into a release. We can safely drop it.

compat: drop 4.3 "new troubles" report code This is supported natively for Mercurial 4.4 and above.

compat: drop 4.3 wrapping of createmarker for effect flag support This is supported natively in 4.4 and above.

compat: drop 4.3 monkey patch of `copies.mergecopies`

compat: drop 4.3 compatiblity code for `obsutil.allprecursors` function

compat: drop 4.3 compatiblity code for `ui.edit` method

compat: drop 4.3 compatiblity code for 'successors' revset

compat: drop 4.3 compatiblity code for 'precursors' revset

compat: drop Mercurial 4.3 support for exthelper The last release email noted plans to drop 4.3 and 4.4 support in the next feature release. I'd like to move this code into core, and dropping this should allow the class to be copied in unmodified.

compat: update metadata about minimum ag version

next: solve the issue of `next` get confused by split This patch solve a problem of next command which get confused by split. Let me describe how it was getting confused: Initial state of repo: A---B---C After splitting B to (B1,B2): @ A---B1---B2 \ ---B---C X * (note: C is orphan; checkedout to B1) Lets make an amend on B1: @ B1' / A---B1---B2 \ X * \ ---B---C X * Now, if run `hg next` (--evolve is True by default now): $ it would give you choice to choose from B2 and C thinking that C could also be a possbile children for B1, instead of stablizing B2 on B1. I fixed this problem by filtering those aspiring children which can be stablized on one of the aspiring children itself. Changes made in test-prev-next.t shows the changed expected behaviour.

next: add test which shows that `next` get confused by split

cleanup: avoid a Yoda condition

evolve: modify and move the comment to appropriate position

next: update the command description Now, if necessary next command evolve the next child revision by defualt. You don't need to pass the --evolve flag.

next: make next command --evolve by default Before this patch, if we need to evolve to update to the next child, we were suggesting the user to use --evolve flag. This patch make some changes to evolve by default in that conditions. After making next command to evolve by default we have to consider the following points: 1) If we don't need to evolve while updating to the next child: a) And if wdir is dirty, we suggest to use --merge flag b) if wdir is clean, we simply update to next child (if ambiguous, prompt the user to select one) 2) If we need to evolve: a) when wdir is dirty, we suggest the user to use `hg shelve` first, to make wdir clean. As we don't support --merge while evovling. b) when wdir is clean, we evolve the next cset. Changes made in test-prev-next.t reflect the changed behaviour.

safeguard: allow push to succeed (and without warning) with --publish

topic: only add --publish flag to push if it's not already there

safeguard: check auto-publish value before sending listkeys command Always sending listkeys command is just wasteful, let's first make sure users even care about auto-publish behavior.

style: silence another flak8 warning Default should be green now.

oops: backed out changeset df22f010cf24 pushed by error The changeset still fails some test

next: make next command --evolve by default Before this patch, if we need to evolve to update to the next child, we were suggesting the user to use --evolve flag. This patch make some changes to evolve by default in that conditions. After making next command to evolve by default we have to consider the following points: 1) If we don't need to evolve while updating to the next child: a) And if wdir is dirty, we suggest to use --merge flag b) if wdir is clean, we simply update to next child (if ambiguous, prompt the user to select one) 2) If we need to evolve: a) when wdir is dirty, we suggest the user to use `hg shelve` first, to make wdir clean. As we don't support --merge while evovling. b) when wdir is clean, we evolve the next cset. Changes made in test-prev-next.t reflect the changed behaviour.

next: improve the message for dirty working copy with --evolve This will be important soon when --evolve will be the default.

next: use the `hint` argument of bail if changed The argument is availabel since Mercurial 4.1. So let's use it.

prev: use the `hint` argument of bail if changed The argument is availabel since Mercurial 4.1. So let's use it.

next: explicitly specify --no-evolve in the tests The default value for this flag is about to switch.

branching: merge with stable

style: fix various flake8 warning

next: set if we need to evolve or not; also move logic to appropriate pos set the value `needevolve` and also move the logic to find aspiringchildren to appropriate position i.e before we suggest user to use other flags like --merge.

next: move some code to prepare for `next --evolve` by default Moved some values before the block where we suggest the user to use --merge flag. As we would need these values to find if we need to evolve or not before we can suggest user to use `--merge` flag. We will suggest user to use `--merge` flag only when we don't need to evolve. Because currently we don't support both --evolve(going to be True by default) and --merge at the same time.

branching: merge with stable

topic: improve UX by using pager only when we should Before this patch we were using pager in every case of topic command, either it is setting up a new topic or clearing the active topic. After this patch it will not use pager in following commands: $ hg topic [TOPIC] -r [rev] $ hg topic --clear

evolvecmd: _possibledestination() skip visiting empty parent I see that even in the case when "parent == -1" we are visiting that also and find succsets for that and in last -1 is being added in the set of possible destination. Why not skip that?

pick: `hg help` was not showing the full cmd desc of `pick` after running command `hg help` it was not including the second line of command description. I guess it picks the firstline only. So I moved that part in first line of cmd description.

test-compat: merge mercurial-4.4 into mercurial-4.3

test-compat: merge mercurial-4.5 into mercurial-4.4

test-compat: merge mercurial-4.6 into mercurial-4.5

test-compat: merge mercurial-4.7 into mercurial-4.6

test-compat: merge stable into mercurial-4.7

compat: fix fixedcopytracing compatibility with mercurial 4.8 and narrow Mercurial 4.8 version of _fullcopytracing includes some new narrow code. As https://phab.mercurial-scm.org/D3896 has not yet landed, update fixedcopytracing behind a version detection condition. This was spotted by Augie Fackler.

utility: don't translate commit messages in revselection prompt (issue6016) Commit messages should not be translated and should be shown the same. I am not sure if I can add a test for this by creating a commit with non-ascii character in commit message.

evolve: use stack template in --continue case also We use stack indexes (s#) in evolve msging when revs being evolved are in active topic. Before this patch we are not covering the --continue case. So this patch cover the same and use the stack indexes s# even when we continue an interrupted evolve. In this patch I used the same concept (to check if we should use s#) as we are using in cases other than `--continue`. Thanks to PYD for refactoring those lines. Changes in tests also reflects the expected behaviour.

evolve: add test showing currently stackalias fails in --continue case

branching; merge with stable

packaging: mark version as devel version This avoid confusion from people installing from source.

Added tag 8.3.1 for changeset b90422a11a88

packaging: prepare version 8.3.1

test-compat: merge mercurial-4.4 into mercurial-4.3

test-compat: merge mercurial-4.5 into mercurial-4.4

test-compat: merge mercurial-4.6 into mercurial-4.5

test-compat: merge mercurial-4.7 into mercurial-4.6

test-compat: merge stable into mercurial-4.7

evolve: do not try to display stack index of changeset who do not have one

evolve: check if we should use stack template for each item in the stack We should probably just track the current topic and move all the logic in the template itself. However, this is a good first step.

branching: merge stable back into default

test-compat: merge mercurial-4.4 into mercurial-4.3

test-compat: merge mercurial-4.5 into mercurial-4.4

test-compat: merge mercurial-4.6 into mercurial-4.5

test-compat: merge mercurial-4.7 into mercurial-4.6

test-compat: merge stable into mercurial-4.7

branching: revert test outptu change for 4.7

pullbundle: drop erroneous flagging of the command as basic This is a debug command, not a basic part of the UI.

test-compat: backed out changeset ba1e83c841f2 Preparing the branch for 4.8.

test-compat: backed out changeset b0084d599112 Preparing the branch for 4.8.

test-compat: backed out changeset e2d083cc447e Preparing the branch for 4.8.

test-compat: backed out changeset ccbfbeee6cf3 Preparing the branch for 4.8.

test-compat: backed out changeset 7091f7215651 Preparing the branch for 4.8.

test-compat: backed out changeset 8eec9422bd8e Preparing the branch for 4.8.

branching: merge stable into default

branching: start test compatibility branch for Mercurial 4.7 Mercurial 4.8-rc0 have been released

topic: add the changelog argument to branchinfo() 47c03042cd1d in core added a changelog argument to revbranchcache.branchinfo() and we wrap that function in topic extension. So let's add this optional argument in topic extension too.

evolve: adapt for changed @command decorator Core changed the hacky "^command" syntax to be a proper decorator option in fa88170c10bb (help: adding a proper declaration for shortlist/basic commands (API), 2018-10-12). This patch makes evolve compatible with that.

topicidx: stop assigning index number to obsolete changesets Obsolete changeset are no longer part of the stack, the old logic tried to assigns them the index of their successors but this is flacky and error prone. So we stop doing so.

stack-index: add some test featuring content diversion This will be useful to display various issues.

pullbundle: drop erroneous flagging of the command as basic This is a debug command, not a basic part of the UI.

topic: add the changelog argument to branchinfo() 47c03042cd1d in core added a changelog argument to revbranchcache.branchinfo() and we wrap that function in topic extension. So let's add this optional argument in topic extension too.

topics: "prairie" is a not an animal (AFAIK, anyway)

tests: remove a"| head -n 2" that just hides an error code The output from core used to be longer and it made sense, but no longer does. CORE-TEST-OUTPUT-UPDATE: 5199c5b6fd29

evolve: adapt for changed @command decorator Core changed the hacky "^command" syntax to be a proper decorator option in fa88170c10bb (help: adding a proper declaration for shortlist/basic commands (API), 2018-10-12). This patch makes evolve compatible with that.

commands: adjust metavariables as appropriate Apart from looking better in hg help command, these strings are also helpful when generating shell completions programmatically.

branching: merge with default

branching: preserve the expected output on default

evolvecmd: use modern vocabulary in help text for evolve

evolvecmd: mark flags with old instability names as deprecated

packaging: mark as development version

Added tag 8.3.0 for changeset 45d4b49d81d9

packaging: prepare release 8.3.0

pullbundle: document working version and update changelog to mention it

test-compat: merge mercurial-4.4 into mercurial-4.3

test-compat: merge mercurial-4.5 into mercurial-4.4

test-compat: merge mercurial-4.6 into mercurial-4.5

test-compat: merge stable into mercurial-4.6

topic: fix '{topicidx}' keyword registration for mercurial <= 4.5

branching: prepare release 8.3.0

test: other harmless output change These are new test that need adjustment for 4.7 output.

changelog: mention the fix of cache warming on hg strip

test: backed out output changes from changeset 6f391c2a773b

test: backed out test output changes from changeset ef22eef37ecc

test: backed out output changes from changeset 0171a7ceb574

test: backed out output changes from changeset dfa69b5ece87

test: backed out output changes from changeset 01ad46551d7c

prev: also use stack shortcut in output when possible This mirror what `hg evolve` is doing since cc3a0b13ae57.

next: also use stack shortcut in output when possible This mirror what `hg evolve` is doing since cc3a0b13ae57.

branching: merge stable into default

rewind: alias the command to undo This is a step in unifiying rewind and undo and not talk about them as different commands or concepts. If there exists rewind command, that should be an alias of undo and the vice versa. The exact implmentation detail and how the command works etc. is a different thing but there should be just one command.

changelog: add a line about the usage of s# in evolve

caches: no longer warm the cache on strip in "auto" mode Before this patch, a strip would trigger a full and blind cache load that would include the stablerange cache. In repository like Mozilla this was a bit painful. The default "auto" mode now behave has intended.

evolve: use stack alias s# in `hg evolve` msgs As repository grows revision number lose their usefulness. To make it a little better than now, after this patch we will be using short stack aliases introduced by the topic extension (as displayed by `hg stack`). These stack aliases s# will be used only when user has enabled topic extension and his repo's current active topic is same as the revs to be evolved. For now, this patch doesn't cover `continue` case. I will discuss about this with Pierre-Yves David. Maybe we have to store the info in evolvestate that "we were using stack aliases".

evolve: extract displayer to _solveone() I extracted this displayer method to _solveone() so that it would be easy to pass a different template to show evolve msgs. In upcoming patches we will be using stacktemplate in some cases where we want to include stack alias s# in evolve msgs.

utility: add a template alias that will be used in evolve msgs This template use stack alias s# instead of rev number in messages printed when we run `hg evolve`. For example, this template's msgs will look like this: move:[s3] add aaa atop:[s2] add bbb move:[s4] add ccc I will be using this template in upcoming patches.

topic: add topicidx template keyword This keyword will be useful where we want to see the index of the changeset in the stack when we run `hg stack`.

changelog: fix version number

tests: update the test-topic.t with changes in core mercurial Yuya stabilized the -T option and now it's in help by default. CORE-TEST-OUTPUT-UPDATE: 0d703063d0c8

pullbundle: fix style warning

pullbundle: improves management of multiple branch while dealing with a head There were still case where a range was issued too early.

pullbundle: deal with head in a predicable order We use node order to make this repeatable from one clone to another. This is an arbitrary pick.

pullbundle: deal with another special case introduced by arbitrary ranges previous, we dealt with the case where the tippest subrange of the stable parent was too long. Now, we need to deal with the case where it is too small.

pullbundle: compat with Mercurial 4.7

pullbundle: add some information on the cached change This also adds a way to skip "caching" smaller range in the debugcommand (not in the extensions itself).

pullbundle: add data about the cached bundle size and hit

pullbundle: add a command to generate cache hit statistic This new command should help checking that the stable range produce reusable bundle.

pullbundle: point out the output flooding issue from core Right now, we don't have a reliable progress report when using pullbundle.

pullbundle: issue local branches in the right order

pullbundle: delay cache file opening Otherwise we can end-up with a too many file open at the same time.

pullbundle: improve slicing of the lower part of range The previous method could get confuse by merge and overslice. The new method is better at using sticking on power of two boundaries.

stablerange: fix slicing of arbitrary ranges Pull bundle trigger slicing from range with arbitrary initial skip. We have to adjust the current slicing to take this into account.

pullbundle: attempt to save stablerange cache after each computation Since we do not serve the full repository, we use more stablerange starting from a random point. These are probably not loaded yet and should be cached. This is best effort caching, if we can't save them we won't.

pullbundle: add basic extension metadata

pullbundle: add a config option for the cache directory

pullbundle: add an output about the time spent slicing the set This will be useful to monitor overall performance.

pullbundle: add extension documentation

pullbundle: add an extra test showing how cache are reusable We issue a single final pull, most of its item are reusable.

pullbundle: add caching logic We now only generate a bundle once (and store it to disk). If we need it again, we use it directly from disk.

pullbundle: slice pull into multiple ranges We subdivide the set of missing revisions into multiple "range" using the "stable range" data structure. This slicing aims at maximizing the capability of the resulting ranges.

pullbundle: initial commit Right now, we do not do any slicing, but we install the wrapping and the tests.

test: update output to make streamv2 capabilities CORE-TEST-OUTPUT-UPDATE: 4bd6e444c76f

branching: merge with stable

topic: respect preexisting 'topic' value on workingcommitctx Mercurial's core changeset: 6c8ceebce309 revealed an issue. If the creator of a workingcommitctx explicitly assigned a 'topic', it was silently overridden by the current active on. This break import of patch with topic information. We now respect preexisting topic information.

topic: add a compatibility to access transaction's names The attribute was made private in 4024c363cd33.

topic: add a compatibility to access transaction's validator The attribute was made private in ebbba3ba3f66.

evovle: remove redundancy in evolve output Copying the discription of this redundancy issue given by Pierre Yves David: When running `hg evolve` to stabilize orphan changeset output about the currently stabilized changeset is issued. For example: $ hg evolve move:[3] a3 atop:[4] a2 working directory is now at 7c5649f73d11 This output can become quite repetitive when orphan are stabilized atop each other. For example: $ hg evolve --all move:[8] dansk 2! atop:[10] dansk! merging main-file-1 move:[9] dansk 3! atop:[11] dansk 2! In this case it would be smoother to issue: $ hg evolve --all move:[8] dansk 2! atop:[10] dansk! merging main-file-1 move:[9] dansk 3! Since we are moving "dansk 3!" atop the changeset we just stabilized. When adding this be careful that we still want to issue the "atop" message in various cases: 1. first changesets in a stack 2. when the orphan is not stabilized atop previous one 3. when using hg evolve --continue to resume an evolution So, I have made the changes which also respect above listed three points. And changes in tests/test-evovle*.t reflecting the changed behavior.

amend: handle crash before transaction initialization with --patch The `tr` object can be None.

branching: merge with stable

obsdiscovery: adopt to calling convention change Upstream commits 71d83b315778 and abce899c985f changed the calling convention of setdiscovery._takefullsample(). We inspect the signature of the function before calling it so we can pass the proper arguments.

branching: merge stable into default

packaging: mark as development version This avoid confusion when people install version from source.

Added tag 8.2.1 for changeset c6362c4abd69

packagin: prepare release 8.2.1

test-compat: merge mercurial-4.4 into mercurial-4.3

test-compat: merge mercurial-4.5 into mercurial-4.4

test-compat: merge mercurial-4.6 into mercurial-4.5

test-compat: merge stable into mercurial-4.6

branching: merge stable into default

caches: issue both debug and blackbox log message This increase the odd someone looking for issue will find them.

cmdrewrite: rename variable and move it closer to where it is used Renaming to satisfy pep8-naming linter plugin, which complained on this variable because all-uppercase names are reserved for constants, and constants need to be module-level. This variable is not a good choice for moving up to the module level because it's used only in one place. Additionally, let's have it closer to the place it's actually used to potentially avoid an extra call to _() function (however cheap it may be).

changelog: mention the cache improvements

obshashrangecache: be more paranoid in the errors we catch It seems like there error were already covered. Better safe than sorry.

stablerangecache: be more paranoid in the errors we catch It seems like there error were already covered. Better safe than sorry.

stablesortcache: update the variable tracking on-disk state after write Since we updated the on disk content, we should update that value. In practice the object will likely be discarded after the write, but there is nothing wrong in being more correct.

stablesortcache: ignore permission and OS errors when writing This cache is related to the obshashrange one and we update it lazily by default. This can be an issue when pulling locally from a read only repository that was not configured for a more aggressive cache warming. The raised permission error was uncaught and could crash the whole process. Errors during cache update should not block Mercurial operations.

firstmergecache: update the variable tracking on-disk state after write Since we updated the on disk content, we should update that value. In practice the object will likely be discarded after the write, but there is nothing wrong in being more correct.

firstmergecache: ignore permission and OS errors when writing This cache is related to the obshashrange one and we update it lazily by default. This can be an issue when pulling locally from a read only repository that was not configured for a more aggressive cache warming. The raised permission error was uncaught and could crash the whole process. Errors during cache update should not block Mercurial operations.

depthcache: update the variable tracking on disk state after write Since we updated the on disk content, we should update that value. In practice the object will likely be discarded after the write, but there is nothing wrong in being more correct.

depthcache: ignore permission and OS errors when writing This cache is related to the obshashrange one and we update it lazily by default. This can be an issue when pulling locally from a read only repository that was not configured for a more aggressive cache warming. The raised permission error was uncaught and could crash the whole process. Errors during Cache update should not block Mercurial operations.

obscache: update the variable tracking on disk state after write Since we updated the on disk content, we should update that value. In practice the object will likely be discarded after the write, but there is nothing wrong in being more correct.

obscache: ignore permission and OS errors when writing In practice we always update this cache along side a new transaction, with lock taken so permission issues should not happens. However, staying on the safe side is better. Errors during Cache update should not block Mercurial operations.

obshashrange: force reset if many changes If there is many new markers (or the markers affect many changesets) resetting affected changeset can get slow. A common case for that is then changeset and markers are added in different transaction, (eg: when using clone bundle). When we detect such case, we reset the obshashrange cache, rewarming it should be faster. Here many is arbitrarily set to 10 000. The number might get adjusted.

topic: refactor revset.py slightly Checking both kind == 'literal' and topic.startswith('literal:') seems to be redundant, because of how stringutil.stringmatcher() works.

test: update output to match new `hg verify` reports CORE-TEST-OUTPUT-UPDATE: f1186c292d03

test: update output to match new "pull" phase report CORE-TEST-OUTPUT-UPDATE: 89630d0b3e23

test: adapt to change in revert messages order CORE-TEST-OUTPUT-UPDATE: cb70501d8b71

rewind: fix help text (mostly grammar, but also s/precursor/predecessor)

evolve: fix use of "'%d' % ctx" ctx has not been convertible to int since Mercurial commit 38f480502043 (context: remove basectx.__int__ (API), 2018-02-24). I decided to use "%s" instead to match other similar messages in evolvecmd.py. I also added a space to match the other messages.

stablerange: only issue the message once There is a flag to prevent issuing the long load message multiple time. However, it was never set. This is now fixed

changelog: fix version number

branching: merge back with stable

packaing: mark as development version This help avoid confusion when installing in progress work.

Added tag 8.2.0 for changeset 8d8f08245f97

branching: merge back with stable

packaging: prepare new release

test-compat: merge mercurial-4.4 into mercurial-4.3

test-compat: merge mercurial-4.5 into mercurial-4.4

test-compat: merge mercurial-4.6 into mercurial-4.5

test-compat: merge stable into mercurial-4.6

topic: rework topic templatekw declaration to be 4.5 compatible again Mercurial prior to 4.5 does not have a requires keyword, so we need a compatibility layer.

test-compat: merge mercurial-4.4 into mercurial-4.3

test-compat: merge mercurial-4.5 into mercurial-4.4

test-compat: merge mercurial-4.6 into mercurial-4.5

test-compat: merge stable into mercurial-4.6

branching: merge the two stable heads

test-output: backout test output update in 8fcee1e65e29

test-output: backout test output update in 503ae2689df8

topic: update the changelog with the 's#' alias change

topic: update other occurrence of 't0' is 's0' in the code

topic: mention t0 in topic documentation

stack: display 's#' instead of 't#' and 'b#' If we want people to use it, we should display it.

stack: clarify s# error message in case of branch A branch might have a lots of changeset, with all of them public.

alias: allow reference through 's#' We keep the 't#' option, but we hope to be able to deprecate it at some point. We move to 's#' because it refers to `stack` a more generic concept. than topic. In addition, this allows to phase out 'b#' that collide with normal hex-node.

topic-revset: get the list of revset through the official localrepo API This is simpler and faster. In addition we tighter the logic to be used in case of empty topic. We do not need to fall back to revision search.

topic-revset: strictly read string Making a clear distinction between string and symbol passed to topic is important when it comes to empty topic. Filtering an empty topic name is harmless. Searching a non existent revision is not.

test: search for empty topic as a string The reduce the ambiguity in the test to help with the next change.

topic: handle ambiguous arguments to topic() revset These arguments can be interpreted as either string or a revset. The decision is made based on existence of topic with such a name. This matches the behavior of branch() revset. The code needs to know all topics that ever existed in the repo, because some commands report "disappearance" of topics after certain operations, using this revset (e.g. via stack.stack or repo.revs).

topic: handle revsets passed to topic() revset This handles cases when the argument can't be interpreted as a string at all.

topic: handle string argument to topic() revset earlier

topic: prepare to handle non-string arguments to topic() revset

topic: return result early if there are no arguments to topic() revset

obshashrange: enable by default We have been testing this in various placed and specific stress test have been performed. It is time to enable this by default

obshashrange: issue a warning if the cache takes a long time to load The -current- implementation of the obshashrange cache can get slow. This will be fixed eventually, in the mean time we inform people on large repository of what have been going on. Since we are about to enable obshashrange by default, we need to make sure nobody will have an (unexplained) bad surprise.

pick: fix last references of grab in code Grab was still referenced on comments and variable names.

pick: replace `hg grab` invocation in test with `hg pick`

pick: replace "grabstate" with "pickstate"

pick: use "pick" as operation name in obsolescence markers

pick: use "pick" instead of "grab" for various internal API This should not much user visible impact but this is cleaner.

pick: update command output to mention pick instead of grab

pick: rename the grab command to pick Grab is ambiguous with "graft", "pick" match the associated verb in histedit. The "grab" name is preserved as an alias for now.

grab: prefix function with 'cmd' as for the other command in the extension

readme: document the purpose of the evolve extension Another place for the explanation to live.

prune: replace --biject with --pair People have been complaining about `--biject` being to obscure for a while. A recent discussion with Pulkit Goyal and Dan Villiom Podlaski Christiansen produce some new concrete proposal that we liked. We eventually settle on using `--pair`. The `--biject` flag will be kept as an alias for a while.

topic: check that repo.currenttopic is set before using it as an argument repo.currenttopic can be an empty string, which means there's no current topic set. This fact can be used to short-circuit some checks. Also, this avoids passing an empty string as an argument to topic() revset.

test: update output after merge message change CORE-TEST-OUTPUT-UPDATE: f785073f792c CORE-TEST-OUTPUT-UPDATE: a3fd84f4fb38

readme: document the tag we use on output change

branching: merge with stable

packaging: fix debian version numbers

packaging: fix debian version numbers

branching: merge stable into default

packaging: mark in progress work as development version This avoid confusion when people install development version.

Added tag 8.1.2 for changeset f1cde4c97806

packagin: prepare version 8.1.2

obshashrange: force obshashrange invalidation by bumping schema The fix in 8.1.1 deserve recomputing the obs hash range cache. So we bump its schema version to declare the older cache invalid. The other caches, including the expensive stablerange are unaffected.

test-compat: merge mercurial-4.4 into mercurial-4.3

test-compat: merge mercurial-4.5 into mercurial-4.4

test-compat: merge mercurial-4.6 into mercurial-4.5

test-compat: merge stable into mercurial-4.6

branching: merge with stable

changelog: mention the database robutness fix in the changelog

sqlcache: also catch malformed database error This is apparently another way for sqlite to fail at concurrency.

stablerange: rework saving logic to be clearer and safer

obshashrange: rework saving branching to be clearer and safer

sqlcache: protect read query too Some error (like locked database) can even happens when doing readonly operation. So we protect them too. At that point, it seems that pysqlite3 is not the right tool for this job.

sqlite: fast path when nothing to save If we have no new data to save, we should try to do anything. Doing something concurrently is hard enough. If we can avoid doing it at all that would be great.

sqlcache: initialize meta table last It turns out that pysqlite commit after each table creation. This commit is an hopeless attemps to make the caches more concurrent friendly.

obshashrange: always check in base cachekey against the recorded one Ignoring the check in the none case is just disaster waiting to happens.

stablerangecache: fix output in the drifted case The node were not hex, and the field had different order in the pair. This is now fixed.

sqlcache: also ignore integrity error integrity error can happens when multiple client tries to warm similar cache. Given we leave multiple hole in the cache that can be warmed by anyone, this is harder to avoid that we would like. We simply catch the error in this case. Someone will warm the missing entry later.

sqlcache: cache OperationError when saving This is a cache, so we should not crash when trying to operate on it. OperationError can happens when the database is seen as readonly or locked.

sqlcache: passe better connection option These new options should help with handling transaction consistency and database access on high load.

contrib: introduce a small script to stress tests obsolescence exchange Tester have been reporting some error that seems to originate from concurrent access/update to the cache. This script will be useful to reproduce these situations locally.

changelog: mention the database robutness fix in the changelog