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).

templater: handle None returned from templatedir() My recent 91aa9bba3dc9 (templater: make templatepaths() return a single path, or None, 2020-07-21) didn't account for the fact that `templatedir()` returns `None` in frozen binaries. That is ironic, since the reason I'm working on this is to add support for built-in mapfiles in frozen binaries. This patch updates the callers to handle the `None` case. It's somewhat ugly, but I will have to revisit this soon anyway, since my goal is to make all callers handle that case by trying to read the map file using the resources API instead. Differential Revision: https://phab.mercurial-scm.org/D8810

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

templater: do search for include of unqualified builtin outside of config code Commit 081b08e4ea13 (templater: look for mapfiles in template paths, 2015-05-15) added support for using things like `%include map-cmdline.default` to include built-in map files without using a valid path to them. This patch rewrites that support by moving it into `_readmapfile()` so it can later be adapted for reading from a non-file resource. Differential Revision: https://phab.mercurial-scm.org/D8792

templater: switch to lower-level config.parse() in _readmapfile() I hope to modify this code to also work with resources loaded from memory (for PyOxidizer support). For that, we'll need to handle the lookup of relative `%include` path (not joined with the base directory of the containing file). This patch prepares for that by using `config.parse()` instead of `config.read()`, since the latter expects a file in the file system. As it happens, this change also lets us clean up the `config` class to not need the `_includepaths` field. That will happen next. Differential Revision: https://phab.mercurial-scm.org/D8791

config: pass both relative and absolute paths to `include` callback The `include` callback is responsible for loading configs from `%include` statements. The callback currently gets passed the absolute path [1] to the config to read. That is created by joining the dirname of the file that contains the `%include` statement. For PyOxidizer support, I'm trying to reduce dependence on paths. This patch helps with that by passing the relative path found in the `%include` statement (but with username expansion, etc.) to the `include` callback. It also turns out that the existing callers can easily adapt to using the relative path. Coming patches will clean that up and then we'll remove the absolute path from the callback. [1] The "absolute path" bit is a bit of a lie -- it's going to be an absolute path if the path that was passed into `config.parse()` was absolute. Differential Revision: https://phab.mercurial-scm.org/D8790

tests: add test for bad template %include and __base__ It doesn't seem like we had any tests for `%include non-existent` and `__base__ = non-existent`. The latter raises an error while the former ignores the include. We should probably also make the former an error. Differential Revision: https://phab.mercurial-scm.org/D8798

tests: add test for unqualified include of a built-in template map The feature was added in 081b08e4ea13 (templater: look for mapfiles in template paths, 2015-05-15), but no test case was added. Differential Revision: https://phab.mercurial-scm.org/D8797

merge with stable

commitctx: extract _filecommit too This function is exclusively used in `commitctx`. So we should extract it too for consistency and to reduce the `localrepo` bloat. This is part of a larger refactoring/cleanup of the commitctx code to clarify and augment the logic gathering metadata useful for copy tracing. The current code is a tad too long and entangled to make such update easy. Differential Revision: https://phab.mercurial-scm.org/D8710

commitctx: extract the function in a dedicated module the function have few callers (< 15) is quite long a mostly independent from the repository itself. It seems like a good candidate to reduce the bloatness of the localrepository class. Extracting it will help us cleaning the code up and splitting it into more reasonable-size function. We don't use a copy trick because the amount of code extract is quite small (<5%) and the de-indent means every single line change anyway. So this is not deemed valuable to do so. This is part of a larger refactoring/cleanup of the commitctx code to clarify and augment the logic gathering metadata useful for copy tracing. The current code is a tad too long and entangled to make such update easy. Differential Revision: https://phab.mercurial-scm.org/D8709

commitctx: document a fast path in _filecommit This block cut off a lot of logic, documenting the why and how seems useful to future reader. This is part of a larger refactoring/cleanup of the commitctx code to clarify and augment the logic gathering metadata useful for copy tracing. The current code is a tad too long and entangled to make such update easy. We start with easy and small cleanup. Differential Revision: https://phab.mercurial-scm.org/D8700

hgweb: simplify now that we always have a single path Both `templatedir()` and `web.templatepath` are now always a single path (or None). Differential Revision: https://phab.mercurial-scm.org/D8788

templater: simplify stylemap() now that templatedir() returns a single path Differential Revision: https://phab.mercurial-scm.org/D8787

templater: make templatepaths() return a single path, or None The function returns either a singleton list or an empty list, so it makes more sense to return a value or None. The plural in the name also doesn't make sense, but `templatepath()` is already taken, so I renamed it to `templatedir()` instead. Differential Revision: https://phab.mercurial-scm.org/D8786

templater: simplify templatepaths() to avoid iterating a singleton list The function iterates over a hard-coded list of one element since d844e220792a (templater: don't search randomly for templates - trust util.datapath, 2014-09-28). Differential Revision: https://phab.mercurial-scm.org/D8785

commitctx: move a tiny else clause above the very long one This is simple to have all the simple case unfold before the 100+ line one. Otherwise it is hard to relate the `else` to the initial conditionnal. This is part of a larger refactoring/cleanup of the commitctx code to clarify and augment the logic gathering metadata useful for copy tracing. The current code is a tad too long and entangled to make such update easy. We start with easy and small cleanup. Differential Revision: https://phab.mercurial-scm.org/D8708

commitctx: consider removed as touched This achieve the same result with clearer code. This is part of a larger refactoring/cleanup of the commitctx code to clarify and augment the logic gathering metadata useful for copy tracing. The current code is a tad too long and entangled to make such update easy. We start with easy and small cleanup. Differential Revision: https://phab.mercurial-scm.org/D8707

commitctx: rename "changed" to touched The variable contains content that are both added and modified. "changed" could be confused with "modified" only, so we pick a less ambiguous naming that will help with more unification. For example, it would make sense to shove the "removed" item in there since this is how the variable is used. This is part of a larger refactoring/cleanup of the commitctx code to clarify and augment the logic gathering metadata useful for copy tracing. The current code is a tad too long and entangled to make such update easy. We start with easy and small cleanup. Differential Revision: https://phab.mercurial-scm.org/D8706

commitctx: more filesremoved assignment closer to removed computation A small change that makes the code flow clearer. This is part of a larger refactoring/cleanup of the commitctx code to clarify and augment the logic gathering metadata useful for copy tracing. The current code is a tad too long and entangled to make such update easy. We start with easy and small cleanup. Differential Revision: https://phab.mercurial-scm.org/D8704

commitctx: compute files added from _filecommit returns When possible, lets avoid recomputing the same information again. This is part of a larger refactoring/cleanup of the commitctx code to clarify and augment the logic gathering metadata useful for copy tracing. The current code is a tad too long and entangled to make such update easy. We start with easy and small cleanup. Differential Revision: https://phab.mercurial-scm.org/D8703

commitctx: return "touched" status from _filecommit Instead of mutating a list passed in argument, we simply return the information from the `_filecommit` function. This make for a cleaner API and allow for richer information to be returned. That richer information will be used in the next commit to avoid duplicated computation. This is part of a larger refactoring/cleanup of the commitctx code to clarify and augment the logic gathering metadata useful for copy tracing. The current code is a tad too long and entangled to make such update easy. We start with easy and small cleanup. Differential Revision: https://phab.mercurial-scm.org/D8702

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.