pep-517: remove the `build-backend` key to allow for users to build extensions PEP 517 changed how projects should define their build dependencies. The presence of a `pyproject.toml` file changes the behavior of `pip` to conform to PEP 517. Since we haven't updated the `setup.py` file yet (I'm not even sure we're not an edge case which would make it harder/impossible, I have not tried yet), this is a workaround to issue6589: it allows users to pass `--no-use-pep517` to pip to revert to the legacy installer. The build backend is simply assumed to be the `build_meta:__legacy__` one. Differential Revision: https://phab.mercurial-scm.org/D11393

tests: make removeemptydirs more portable The behavior of pwd change from one system (i.e. default shell/sh) from one system to another so we use fuzzy matching for the output we do not care and we use narrower command call to highlight what we are trying to test here. Differential Revision: https://phab.mercurial-scm.org/D11449

check-code: make it possible to ignore the PWD check in some situation The check-code script does not motivate the banning of $PWD so I am not sure what it is about. And I will needs to use the env variable in the next tests. So I am adding a way to make an exception. Differential Revision: https://phab.mercurial-scm.org/D11450

relnotes: update next Differential Revision: https://phab.mercurial-scm.org/D11448

relnotes: add release notes for 5.9 Differential Revision: https://phab.mercurial-scm.org/D11447

narrow: fix commits of empty files The problem is that when committing a new file with empty contents (or in general empty file with filelog p1 = -1), hg commit with narrow doesn't create a filelog revision at all, which causes failures in further commands. The problem seems to be that: - hg thinks that instead of creating a new filelog revision, it can use the filelog's p1 (the nullrev) - because it thinks the file contents is the same in that revision and in p1 - because `narrowfilelog.cmp(nullrev, b'')` is True (unlike with `filelog.cmp`) It's not clear to me which `cmp` behaves better. But I think it makes sense to change the commit code to not to "reuse" the null rev when adding an empty file with filelog p1 == filelog p2 == -1. This is consistent with never writing the null rev in the manifest, which `hg verify` claims is an invariant: ``` inside/c@4: manifest refers to unknown revision 000000000000 ``` Differential Revision: https://phab.mercurial-scm.org/D11400

narrow: show repo corruption when commiting empty files Differential Revision: https://phab.mercurial-scm.org/D11399

filemerge: be more strict when detecting conflict markers, add `|` markers I received a user complaint about detecting a line that contained 78 `=` followed by `*/` as a conflict marker. We'll never generate that, we generate 7 identical characters and either the end of the line, or a space. Let's be explicit about detecting exactly what we produce to reduce the chances of a false positive. While we're here, add `|||||||` as a detected conflict marker (generated with the `keep-merge3` style conflicts). Differential Revision: https://phab.mercurial-scm.org/D11391

Added signature for changeset 86a60679cf61

Added tag 5.9.1 for changeset 86a60679cf61