exchange: backout changeset c26335fa4225 Changeset c26335fa4225 has good intends but introduce significant behavior regressions for multiple important cases. In short there are many case where push would have caught instability creation/propagation that are no longer covered. These behavior have been covered for many years and even if some related case are not currently caught, the covered one should not be regressed. The next four changesets introduce tests for some of these cases. However we could produce many more tests cases since the area is wide and they are many possible combination. (And we should cover them when getting back to this issue) Since 5.5 is one week away, the most reasonable approach seems to back this out while we devise a new way to move forward that preserve the current behavior, catch more issues and also improves the situation that c26335fa4225 target. In addition to the behavior change, c26335fa4225 also introduced output changes. These output changes does not requires a backout per-se, but are part of the same changeset. However they come with a couple of issues that also requires attention: 1) the bulk of the error message have been shoehorned into a multiple line abort message. This seems quite different from what we usually do. The abort message should be a compact and efficient message, with extra details being issued as normal error output beforehand. (with --verbose/--quiet) support. 2) the current output is unbounded, so if there is many (tens, hundreds, thousands, …) of unstable/obsolete changeset involved in the push, the output can quickly become a scary and un-usuable wall of text. So we need some limitation here (same as we have with the remote head list that says A, B , C and # others).

tests: glob 'mercurial.error' in test-phases.t On python 2 with chg, `mercurial.error` is omitted while printing error. On other cases it's there in error message. I did tried to understand what might be the cause was unable to find one on quick skim through the code.

infinitepush: remove unused import to tempfile Spotted by test-check-pyflakes.t

relnotes: add release notes for relevant changes I did since the 5.4 release After having written the notes, I realized that the added points contain changes that were part of the 5.4.2 release. These were never included in any release notes. I asked on IRC whether they should be added to this file or not, but didn’t get an answer to this question. If they should not be added here, I can remove them. Differential Revision: https://phab.mercurial-scm.org/D8770

relnotes: fix indentation The two points were written by me. For some reason, I missed the fact that the other points were indented by one space. Differential Revision: https://phab.mercurial-scm.org/D8769

relnotes: make spacing before new section consistent Differential Revision: https://phab.mercurial-scm.org/D8768

tests: make one of 'worker process exited' line option in test-chg.t I investigated and it seems like under high load, the child process is already there and no new fork was created. Hence the process was not closed.

infinitepush: fix `{get,put}_args` formatting on Python 3 Calling `.format()` on a byte-string does not work, thus causing an exception on Python 3. This commit adds a function to paper over the difference. Differential Revision: https://phab.mercurial-scm.org/D8781

infinitepush: replace `NamedTemporaryFile` with `pycompat.namedtempfile` Fixes a Python 3 compat error when using the external bundle store. Differential Revision: https://phab.mercurial-scm.org/D8780

chgserver: discard buffered output before restoring fds (issue6207) On Python 3, flush() appears not discarding buffered data on EPIPE, and the buffered data will be carried over to the restored stdout.

hghave: fix possible int('') in has_clang_format()

Added signature for changeset 28163c5de797

Added tag 5.5rc0 for changeset 28163c5de797

merge default into stable for 5.5rc0

phases: move short-lived PyObject pointers to local scope It helps understand which object should be decrefed on goto release.

phases: rename variable used for owned dict of phasesets The phaseroots variable is used for two different objects: borrowed set and owned dict of sets. It's hard to track which object should have to be decrefed on error return.

phases: leverage Py_BuildValue() to build PyInt and steal PyObject "N" means "O" without incref, so we can just return the created tuple.

phases: make sure an exception should be set on error return This should never happen, but the code looks wrong without PyErr_Set*() call.

phases: fix clang-format error

phases: fix error return with no exception from computephases() PySet_Check() does not set an exception.

osutil: fix excessive decref on tuple creation failure in listdir() The stat object would be freed on error for the same reason as the previous patch. makestat() can be inlined, but this patch doesn't change it. https://github.com/python/cpython/blob/2.7/Python/modsupport.c#L292 The __APPLE__ code is untested.

revlog: fix excessive decref on tuple creation failure in parse_index2() Since Py_BuildValue() steals the ownership of "N" arguments, these objects would already be freed if Py_BuildValue() returned NULL. https://github.com/python/cpython/blob/2.7/Python/modsupport.c#L292

tests: correctly match clang-format version >= 10 The previous regex only matched the first digit, so "clang-format version 10.0.0" was recognized as version 1.

state: support validated declaration of nested unfinished ops This enables extensions to define commands that delgate to rebase, evolve, etc. one or more times to also have their own unfinished states for the full sequence of operations without monkey-patching _unfinishedstates. Differential Revision: https://phab.mercurial-scm.org/D8714

absorb: improve message for the case when changeset became empty In changeset f55099982bc5, I introduced the message "became empty and became ...", which I was never very happy with. Raphaël Gomès suggested "became empty as ...". That sounds much nicer. Differential Revision: https://phab.mercurial-scm.org/D8765

relnotes: extend Differential Revision: https://phab.mercurial-scm.org/D8767

merge: document return values of manifestmerge() and calculateupdates() In future patches, I want to add one more return value which represents information which needs to stored and used at commit time. Differential Revision: https://phab.mercurial-scm.org/D8741

merge: refactor code to advise fsmonitor in separate function merge.update() is quite hard to understand, found this an easy win. The end goal is to have better organized merge and mergestate handling and then fix some related bugs. Differential Revision: https://phab.mercurial-scm.org/D8740

mergestate: document what mergestate._results is for Understanding that dict is important for understanding how mergestate is performing operations on dirstate. Differential Revision: https://phab.mercurial-scm.org/D8739

mergestate: rename a helpless variable name to bit helpful one The old variable name `r` makes it ~impossible to understand what does it mean. One can only understand that after going to callers and hoping that its documented there. I also documented return value of the function involved while I was there. Differential Revision: https://phab.mercurial-scm.org/D8738

mergestate: remove unnecessary recordactions() from mergestate class This function is updating dirstate which sounds like not something which a method on mergestate class should do. Also this just calls another function. Lets directly call that function and remove this reducing mergestate responsibility a bit. There was single caller which is updated. Differential Revision: https://phab.mercurial-scm.org/D8737

mergestate: document mergestate records in an organized way This makes clear which mergestate record is used for what and group them based on how they are used right now. Differential Revision: https://phab.mercurial-scm.org/D8719

mergestate: remove unused unsupported related mergestate records I tried to find users of this but was unable to find. Seems like RECORD_OVERRIDE is doing for what they were used before. Differential Revision: https://phab.mercurial-scm.org/D8718

mergestate: add comments about couple of record types and minor reorder I am trying to divide the records into certain groups and then have dedicated objects for them. Taking baby steps in that direction. Differential Revision: https://phab.mercurial-scm.org/D8717

mergestate: remove unrequired RECORD_RESOLVED_OTHER record This was introduced in last cycle however while working on refactoring mergestate, I realized it's unncessary. This will break users who did a merge using previous version, did this kind of storage and before commiting updated the mercurial version. Differential Revision: https://phab.mercurial-scm.org/D8716

mergestate: rename addpath() -> addpathonflict() to prevent confusion addpath() seems to imply that we are adding a new path/entry to the mergestate. Differential Revision: https://phab.mercurial-scm.org/D8715

windows: don’t set `softspace` attribute in `winstdout` Python 2 file objects have the `softspace` attribute (https://docs.python.org/2/library/stdtypes.html#file.softspace), which is used by the print statement to track its internal state. The documentation demands from file-like objects only that the attribute is writable and initialized to 0. Method `file.write()` sets it to 0, but this is not documented. Historically, sys.stdout was replaced by an instance of the `winstdout` class, so it needed to behave exactly the same (the softspace fix was introduced in 705278e70457). Nowadays we don’t replace sys.stdout and don’t use the print statement on `winstdout` instances, so we can safely drop it.

windows: handle file-like objects without isatty() method Copying the function is not nice, but moving around stuff to avoid the circular import didn’t seem to be worth the effort.

exchange: check actually missing revs for obsolete / unstable revs (issue6372) The previous code was using `outgoing.ancestorsof`, which was originally called `outgoing.missingheads` although not containing the missing heads. This confusion was probably the reason why the buggy code was written. The actually outgoing changesets are stored in `outgoing.missing`. By checking all outgoing changesets, we avoid the problem and can show the list of all obsolete or unstable changesets, which is more helpful for the user.

tests: test that push doesn’t complain about unstable changesets if no changes When there’re no outgoing changes, push doesn’t complain about unstable changesets. There is currently a bug (see issue6372) that causes that there is an abort on push when the outgoing changes contain another changeset even if that is not obsolete or unstable. A test case and fix for that is sent in the next patch.

diff: move no-eol text constant to a common location Differential Revision: https://phab.mercurial-scm.org/D8763

revert: fix interactive reverting of end-of-file newline changes The chunk reversal used by `revert -i` in Curses mode was not taking this case into account. Differential Revision: https://phab.mercurial-scm.org/D8762

phabricator: clarify the multiple heads/roots error message for phabsend As pointed out in followup feedback for c1c922391314, it's confusing to say there are multiple items and then only list one. Differential Revision: https://phab.mercurial-scm.org/D8761

error: normalize "unresolved conflicts" error messages with a custom class Differential Revision: https://phab.mercurial-scm.org/D8713

error: unify the error message formats for 'rebase' and 'unshelve' Differential Revision: https://phab.mercurial-scm.org/D8730

debugcommands: create new debugantivirusrunning command This writes the EICAR test file to .hg/cache, in an attempt to trigger an AV scanner's scanning engine. This should let us (in theory) detect some cases when a user's slowness is a result of AV scanning. Differential Revision: https://phab.mercurial-scm.org/D8353

windows: always work around EINVAL in case of broken pipe for stdout / stderr In 29a905fe23ae, I missed the fact that the `winstdout` class works around two unrelated bugs (size limit when writing to consoles and EINVAL in case of broken pipe) and that the latter bug happens even when no console is involved. When writing a test for this, I realized that the same problem applies to stderr, so I applied the workaround for EINVAL to both stdout and stderr. The size limit is worked around in the same case as before (consoles on Windows on Python 2). For that, I changed the `winstdout` class.

tests: make pipes / PTYs non-inheritable in test-stdio.py A following patch requires that to test closing the receiving end of the pipe / PTYs. Even for existing tests, it might be safer to make the lifetime of the pipes / PTYs as short as possible.

util: enhance `nullcontextmanager` to be able to return __enter__ result This makes its functionality and signature equivalent to Python 3.7’s contextlib.nullcontext().

windows: augment docstring of `winstdout` class

discovery: change users of `outgoing.missingheads` to `outgoing.ancestorsof` The attribute `missingheads` was recently renamed to `ancestorsof`, as it, despite the old name, doesn’t contain the missing heads but the changesets that were requested (including ancestors) for the outgoing operation. Changing all the users enables to print a warning if the old name is used. There is a good chance that some of the users are buggy because of the old name. Changing them to use the new name makes it more obvious that they are buggy. All users need to be reviewed for bugs. When sending patches for fixing them, the change will be more obvious without having to explain again and again the discrepancy of the old attribute name and what it actually contained.

discovery: fix docstring of `outgoing` class Also, introduce a more correct name `ancestorsof` for what was named `missingheads` before. For now, we just forward `ancestorsof` to `missingheads` until all users are changed. There were some mistakes in the old docstring / name: * `missingheads` (new name: `ancestorsof`) contains the revs whose ancestors are included in the outgoing operation. It may contain non-head revs and revs which are already on the remote, so the name "missingheads" is wrong in two ways. * `missing` contains only ancestors of `missingheads`, so not *all nodes* present in local but not in remote. * `common` might not contain all common revs, e.g. not some that are not an ancestor of `missingheads`. It seems like the misleading name have fostered an actual bug (issue6372), where `outgoing.missingheads` was used assuming that it contains the heads of the missing changesets.

discovery: weaken claim about returned common heads if ancestorsof are given As the test case shows, the claim is not true in general.

cext: remove unused variables Differential Revision: https://phab.mercurial-scm.org/D8760

phabupdate: allow resigning from revisions Differential Revision: https://phab.mercurial-scm.org/D8756

phabupdate: allow revisions to be commandeered Differential Revision: https://phab.mercurial-scm.org/D8755

phabupdate: allow revisions to be reopened Differential Revision: https://phab.mercurial-scm.org/D8754

phabupdate: allow revisions to be marked "closed" (Yes, the transaction name and final state in the comment here do differ.) Differential Revision: https://phab.mercurial-scm.org/D8753

phabupdate: allows revisions to be marked "request-review" Differential Revision: https://phab.mercurial-scm.org/D8752

phabupdate: allow revisions to be marked with "plan changes" Differential Revision: https://phab.mercurial-scm.org/D8751