Mercurial: Changelog

setup: avoid attempting to invoke the system-wide hg.exe on Windows On Windows, the executable in the current directory gets priority over anything in $PATH (both for cmd.exe and MSYS). That means, the former code was launching the local hg.exe instead of the system-wide one, if it was previously built. If that failed, it then fell back to the local hg code, but run through python.exe. I'm not sure what it is about ef7119cd4965, but that started throwing up a messagebox that python37.dll couldn't be loaded. (And indeed, python37 is not in $PATH by default.) Invoking the local hg via the current python avoids that.

delta: ignore base whose chains already don't match expectations If we know the existing chain does not match our criteria, there is no point to build a delta to append. This is especially useful when dealing with a full text much smaller than its parent. In that case, the parent chain is probably already too large. example affected manifest write before: 1.421005s after: 0.815520s (-42%)

delta: exclude base candidate much smaller than the target If a revision's full text is that much bigger than a base candidate full text, we no longer consider that candidate. This solves a pathological case we encountered on a very specify repository. It contains a long series of changesets with a very small manifest (one file) co-existing with others changesets using a very large manifest. Without this filtering, we ended up considering a large number of tiny full snapshots as a potential base. It resulted in very large delta (the size of the full text) and mercurial spending 99% of its time compressing these deltas. The timing of a commit moved from about 400s to about 10s (still slow, but not ridiculously slow).

perfrevflogwrite: clear revlog cache between each write We want to measure write time from a cold cache (similar to commit). So we need to clear the cache to prevent computation from rev N-1 to interfere with rev N.

py3: enable legacy stdio mode in exewrapper This drops the test failure count from 166 to 117. The failures were typically in the form of `hg serve -d` spawning but crashing immediately, leaving clients with "bad http status" lines, connection refusals, and so forth. The underlying message on the server side was either "OSError: [WinError 6] The handle is invalid" or "OSError: [WinError 1] Incorrect function". Additionally, no output was rendered if the pager was activated. Thanks to Yuya for diagnosing the problem. The failure count drops to 107 when PYTHONLEGACYWINDOWSSTDIO=1 is defined in the environment. These failures seem to revolve around the dummyssh server process, and dumbhttp.py. So I'll probably add that to the test runner. One small regression here (only in py3) is that if hg.exe is already built, a messagebox appears when building it again saying that python37.dll can't be loaded. Python3 isn't in PATH by default, and setup.py tries running bare `hg` first. But MSYS prepends '.' to PATH, so it runs the local hg, but can't find the library. It falls back to the python used to invoke setup.py, so ultimately it works. I'm not sure if it's better to strip '.' from PATH or just skip right to `sys.executable hg` on Windows. Also, something seems to be wrong with run-tests._usecorrectpython(). I accidentially left off the 'PYTHON="py -3"' when building (thus making py2 stuff), and yet managed to invoke run-tests.py with "py -3". (And that only had 67 failures.)

run-tests: alias hg to hg.exe on Windows To enable legacy stdio mode on Windows, hg.exe needs to be updated. But before that, we actually have to use it when running the tests. I *think* what was happening before was when MSYS invoked `hg`, it looked at the shbang line and ran python.exe found there. The test harness must be updating $PATH to include the python used to launch it, and therefore it ran py3. As a side note, this also fixed `py -3 run-tests.py` (without --local), which complained about the space in the shbang line before this. This should also help in WSL, because the explicit '.exe' is needed to invoke a Windows app instead of invoking the Linux app. I have no idea why this change capitalizes Lib in the tests, as it was previously lowercase for both py2 and py3.