merge with stable

merge: respect parents order when using `graft` on a merge, this time for real See a4ca0610c754. potherp1 is a boolean variable that means "pother is ctx.p1", and parents is naturally [ctx.p1, ctx.p2]. pctx is always removed from parents, so if pctx is parents[0], then we end up using parents[1] as pother. To be true to its name, potherp1 should then be True only when pctx is at parents[1].

py3: don't double-convert "opts" to bytes The "opts" are already converted to bytes at the beginning of the function. Doing it twice results in a crash, which makes test-uncommit.t fail. The extra call was added recently, in ff1ff2aae132 (uncommit: add support to modify the commit message and date, 2019-09-07). test-uncommit.t passes again after this patch. Differential Revision: https://phab.mercurial-scm.org/D6864

py3: byte-prefix sanitisation regexes in phabricator.py So it doesn't die with "TypeError: cannot use a string pattern on a bytes-like object". Differential Revision: https://phab.mercurial-scm.org/D6863

py3: pass a bytestring into querydrev instead of a string that'll TypeError This was a regression I introduced in c19d259fd6ad. When the string gets to the memoryview in _tokenize under py3 it'll die. Differential Revision: https://phab.mercurial-scm.org/D6869

py3: add test demonstrating TypeError when phabsending skips unchanged commits Skipping can currently only happen with `--no-amend`, so this isn't a usual configuration. Differential Revision: https://phab.mercurial-scm.org/D6868

tests: clean up built binaries after running test-fuzz-targets.t Most users won't notice a change here because they won't have the fuzzer infra, but by good fortune my workstation has the required bits and keeps leaving the fuzzer binaries around. Differential Revision: https://phab.mercurial-scm.org/D6862

fastannotate: remove support for flock() locking We've seen enough weirdness in CI with flock for remotefilelog that I'm now of the opinion we should just stop using flock() everywhere until someone has a concrete need for the extra performance *and* a way to only use it when safe (even if that's just default-to-off.) Differential Revision: https://phab.mercurial-scm.org/D6861

remotefilelog: remove dead code for using flock() for locking Differential Revision: https://phab.mercurial-scm.org/D6860

narrow: add option for automatically removing unused includes It's been a somewhat common request among our users to have Mercurial automatically pick includes to remove. This patch adds an option for that: `hg tracked --auto-remove-includes`. I'm not sure if this is the right name and semantics for it. Perhaps the feature should also add excludes of large subdirectories even if other files in the include are needed? Narrow clones are experimental, so we can change the name and/or semantics later if necessary. Differential Revision: https://phab.mercurial-scm.org/D6848

narrow: don't hexify paths and double-hexify known nodes on wire (BC) It isn't obvious, but wireprototypes.encodelist() is meant only for binary nodeids. So when we used it for encoding hex nodeids and paths, the encoded result was surprising and hard to read. This patch changes the encoding to make the list of paths a comma-separated list and the list of common nodes to be a encodelist()-encoded list of binary nodeids (so the result is just singly-hexified nodeids). This is clearly a breaking change, but the feature is experimental and we're not aware of anyone running a server using this command yet. Differential Revision: https://phab.mercurial-scm.org/D6851

remotefilelog: replace repack lock to solve race condition 2c74337e6483 reduced the probability of race-conditions when starting background repack and prefetch and we saw the difference in our CI instance with all failures disappearing except one where one call to waitonrepack seems to returns too early. I'm not sure what exactly goes wrong but I realized that while the prefetch operation uses a standard Mercurial lock, the repack operation is using a custom lock based on `fcntl.flock` on available platforms. As `extutil.flock` fallback on traditional Mercurial locks on other platforms and the tests are stable on my laptop, our CI environment and GCC112, I'm sending this patch to standardize the behavior across environments. Differential Revision: https://phab.mercurial-scm.org/D6844

perf: add a --stats argument to perfhelper-pathcopies The arguments will display some statisting about the distribution of the value we measure.

perf: add a --stats argument to perfhelper-mergecopies The arguments will display some statistics about the distribution of the value we measure.

archive: add XZ support if built with Python 3

rust-cpython: add sanity check to PySharedState::decrease_leak_count() If decrease_leak_count() were called unnecessarily, there must be a serious bug. It's better to not silently ignore such cases.