fuzz: if the caller of our makefile sets CC and CXX, trust them This should fix the broken fuzzing build, because we've been explicitly using clang++ but are now being given a CXX=afl++, which does extra stuff. Differential Revision: https://phab.mercurial-scm.org/D10066

rhg: Use clap’s support for global CLI arguments By default, clap only accepts app-level arguments (as opposed to sub-command level) to be specified before a sub-command: `rhg -R ./foo log`. Specifying them after would be rejected: `rhg log -R ./foo`. Previously we worked around that by registering global arguments both at the app level and on each sub-command, but that required looking for their value in two places. It turns out that Clap has built-in support for what we want to do, so let’s use it. Also, Clap "settings" turn out to be either global or not too. Let’s make `AllowInvalidUtf8` apply to sub-commands too. Differential Revision: https://phab.mercurial-scm.org/D10080

revlog: add a mechanism to verify expected file position before appending If someone uses `hg debuglocks`, or some non-hg process writes to the .hg directory without respecting the locks, or if the repo's on a networked filesystem, it's possible for the revlog code to write out corrupted data. The form of this corruption can vary depending on what data was written and how that happened. We are in the "networked filesystem" case (though I've had users also do this to themselves with the "`hg debuglocks`" scenario), and most often see this with the changelog. What ends up happening is we produce two items (let's call them rev1 and rev2) in the .i file that have the same linkrev, baserev, and offset into the .d file, while the data in the .d file is appended properly. rev2's compressed_size is accurate for rev2, but when we go to decompress the data in the .d file, we use the offset that's recorded in the index file, which is the same as rev1, and attempt to decompress rev2.compressed_size bytes of rev1's data. This usually does not succeed. :) When using inline data, this also fails, though I haven't investigated why too closely. This shows up as a "patch decode" error. I believe what's happening there is that we're basically ignoring the offset field, getting the data properly, but since baserev != rev, it thinks this is a delta based on rev (instead of a full text) and can't actually apply it as such. For now, I'm going to make this an optional component and default it to entirely off. I may increase the default severity of this in the future, once I've enabled it for my users and we gain more experience with it. Luckily, most of my users have a versioned filesystem and can roll back to before the corruption has been written, it's just a hassle to do so and not everyone knows how (so it's a support burden). Users on other filesystems will not have that luxury, and this can cause them to have a corrupted repository that they are unlikely to know how to resolve, and they'll see this as a data-loss event. Refusing to create the corruption is a much better user experience. This mechanism is not perfect. There may be false-negatives (racy writes that are not detected). There should not be any false-positives (non-racy writes that are detected as such). This is not a mechanism that makes putting a repo on a networked filesystem "safe" or "supported", just *less* likely to cause corruption. Differential Revision: https://phab.mercurial-scm.org/D9952

narrow: remove assertion about working copy being clean The user can always modify the working copy, including while they're running `hg tracked --remove-include <path>`. Nothing really bad happens when they do that, and we already have code for printing a nice warning, so we can safely remove the assertion we had. Differential Revision: https://phab.mercurial-scm.org/D10063

tests: demonstrate assertion error when modifying working copy while narrowing Differential Revision: https://phab.mercurial-scm.org/D10062

ci: test real dependency installation for pip In the past, the pip smoke test inhibited actual dependency installation, but that fails in different environments for setuptools itself. Since it isn't what we actually want to test (which is pip install), allow this to call home, if HGTESTS_ALLOW_NETIO=1 is set in the environment. Differential Revision: https://phab.mercurial-scm.org/D9950

rust: Add some unit tests for parse_byte_size in config Differential Revision: https://phab.mercurial-scm.org/D10022

rust: Move config value parsing functions to a new module Differential Revision: https://phab.mercurial-scm.org/D10021

rhg: Add support for the blackbox extension Only `command` and `commandfinish` events are logged. The `dirty`, `logsource`, `track` and `ignore` configuration items are not supported yet. To indicate commands executed without Python, a `(rust) ` prefix is added in corresponding log messages. Differential Revision: https://phab.mercurial-scm.org/D10012

blackbox: Remove misleading quotes in config example This example previously looked like quotes were part of configuration file syntax, and the parsed `date-format` value was the part inside of them. This is not the case: config syntax only parses quotes in list values. Instead using that config would result in literal quotes being written to `.hg/blackbox.log` as part of the date format. This changes the example to what was probably intended. Differential Revision: https://phab.mercurial-scm.org/D10011

rust: Add a log file rotation utility This is ported to Rust from `mercurial/loggingutil.py`. The "builder" pattern is used to make it visible at call sites what the two numeric parameters mean. In Python they might simply by keyword arguments. Differential Revision: https://phab.mercurial-scm.org/D10010

rust: Add a `ConfigValueParseError` variant to common errors Configuration files are parsed into sections of key/value pairs when they are read, but at that point values are still arbitrary bytes. Only when a value is accessed by various parts of the code do we know its expected type and syntax, so values are parsed at that point. Let’s make a new error type for this latter kind of parsing error, and add a variant to the common `HgError` so that most code can propagate it without much boilerplate. Differential Revision: https://phab.mercurial-scm.org/D10009

rust: Add config parsing support for more value types * Rust `str` (ASCII or UTF-8) * Integer * Byte quantities Differential Revision: https://phab.mercurial-scm.org/D10008

rust: Introduce a get_bytes_from_os_str utility function It does the same as get_bytes_from_path but takes an `OsStr` instead of a `Path`. The implementation is the same so using either ends up correct but the function name suggests it’s not. Differential Revision: https://phab.mercurial-scm.org/D10007

rust: Make `DirstateParents`’s fields typed `Node`s Instead of plain byte arrays. Differential Revision: https://phab.mercurial-scm.org/D10006