revlog: split the content verification of a node into a separate method This will be used by LFS to tune what is skipped. In the future, this could also be used by LFS to indicate which nodes tagged with `skipread` are simply in need of a blob fetch, so that they can be done in a batch later. (Currently, `skipread` also indicates censored data and errors.) Additionally, it could be used to cache the sha1 hash value for each blob so that large blobs don't need to be re-read and re-hashed if they are used by multiple nodes. Differential Revision: https://phab.mercurial-scm.org/D7710

verify: update comment to say that lfs doesn't need fulltext to check renames The reason is that `filelog.renamed()` indirectly calls `filelog.revision()`, which is what accesses the full text. However, LFS wraps `filelog.renamed()` and completely handles the case where an LFS blob is in play by using rawdata. I've got a test to demonstrate that this is the case, and prevent regressions. But the `skipread` flag is set on all lfs revisions when using `--no-lfs`, regardless of whether or not the blobs are local. Just above this, that flag is consulted, causing the rename checks to be skipped. That will need to be loosened up first. Differential Revision: https://phab.mercurial-scm.org/D7709

resourceutil: use `from importlib import resources` Without this patch, we get the following error from when trying to run hg with PyOxidizer: module 'importlib' has no attribute 'resources' I don't know what why that happens, but `from importlib import resources` is the form I would prefer anyway, so let's use that now that the impoort-checker has been fixed. Differential Revision: https://phab.mercurial-scm.org/D7701

import-checker: allow all absolute imports of stdlib modules Before this patch, we didn't allow imports like from importlib import resources (That's the reason I used `import importlib.resources` in D7629.) I think that form is still an absolute import, so I don't think we forbade on purpose. Differential Revision: https://phab.mercurial-scm.org/D7700

help: drop a reference to Windows 9x Differential Revision: https://phab.mercurial-scm.org/D7694

help: clarify that the Windows registry key for hgrc files is systemwide Since there's no version or path info here to distinguish between installations, it is effectively systemwide (unless splitting hairs about the WoW64 registry redirection). Differential Revision: https://phab.mercurial-scm.org/D7693

windows: add a global equivalent to /etc/mercurial for *.rc processing This follows the Unix model of processing this directory immediately after <internal>/*.rc, and prior to the installation relative files. Since the Unix processing supports both a directory and a file (the former overriding the latter), and since %HOME% supports both `*.ini` and `.hgrc` (again, the former overriding the latter), this does too. The Unix file doesn't have a `.` prefix, so it's not used here either. Note that this is the opposite order of processing the exe relative paths. But since it's in agreement with Unix, %HOME% and %USERPROFILE%, it seems reasonable to ignore that. Maybe we can change that and take a BC, because that's something the installer should be controlling, and I can't imagine people having both paths *and* conflicting settings. Differential Revision: https://phab.mercurial-scm.org/D7692

match: normalize `cwd` early By having cwd in absolute form, we won't have to adjust it when passing it to subrepo matchers. This will matter for a coming patch. Differential Revision: https://phab.mercurial-scm.org/D7650

match: make sure `root` argument is always an absolute path (API) The `root` argument should already be an absolute path, but we had tests that passed a relative path. This patch fixes up the tests and adds an assertion. This assumes that `os.path.isabs('/repo')` will be `True` on all platforms we care to run tests on. Augie tested for me that it does work on Windows, so that's good enough for me. Differential Revision: https://phab.mercurial-scm.org/D7649

tests: show that fileset patterns don't work with `fix` when not in repo root Differential Revision: https://phab.mercurial-scm.org/D7569

py3: force bytestr conversion of "reason" in scmutil.callcatch() For instance, reason may be an InvalidURL as shown in added test.

tests-pure: fixing test-parseindex2 This is a followup to D7603 (49fa0b31ee1d) which broke the tests for pure Python implementation. There are two divergences between pure and C implementations: - the pure implementation would accept only -1 as slice end, whereas C accepts both -1 and len(index) - in pure Python, `headrevs` is provided by revlog.py, not by the index. Differential Revision: https://phab.mercurial-scm.org/D7707

rust-index: add a function to convert PyObject index for hg-core Function in hg-core need something implementing the `Graph` trait. Right now, the `hg-cpython` entry points directly turn the PyObject passed as argument into a `cindex::Index`. However, if we start having the option to use an Index in Rust, we need to dispatch between the different possible PyObject we could receive. So move the "duplicate" call into a unified function. When time come. It will be easy to update the logic of all interface when the time come. Differential Revision: https://phab.mercurial-scm.org/D7653

test: extract some generic data and utility from test-rust-ancestor.py We will reuse this for more tests related to revlog index. In pratice this series of changesets add an index implementation provided from Rust and we want to be able to test it. Differential Revision: https://phab.mercurial-scm.org/D7652

patch: fix header.__repr__() to not have `b''` prefixes in file names Spotted by Denis Laxalde. Differential Revision: https://phab.mercurial-scm.org/D7695

cmdutil: allow native string as input to check_at_most_one_arg() We seem to always convert **opts args to use bytes as keys early on in Mercurial core, but I'm not sure we have good reason to do that, and not all extensions do that. It's therefore helpful to be able to pass in a native string to check_at_most_one_arg(). Differential Revision: https://phab.mercurial-scm.org/D7699

cmdutil: return underscore-separate name from check_at_most_one_arg() As noticed by Yuya, when I changed the function (during review) to work with underscore-separated names as input, I forgot to make sure the returned name was also underscore-separated. We don't have any cases where it matters yet, but it should still clearly be fixed. Instead of converting the hyphen-separated value we already have in `previous`, I'm changing it so we convert to the underscore-separated values to be hyphen-separated only when we need to display them. This will also help a coming change where we allow the inputs to native strings instead only bytes. Differential Revision: https://phab.mercurial-scm.org/D7698

rebase: restore i18n of a hint message As noticed by Yuya, I lost the _() call in 71fee4564410 (rebase: use rewriteutil.precheck() instead of reimplementing it, 2019-12-18). Differential Revision: https://phab.mercurial-scm.org/D7697

status: make unresolved files always be in the morestatus structured output We don't know the status of those files, only that they're unresolved, so we don't output the status for those - any code parsing this will have to be tolerant to that. Differential Revision: https://phab.mercurial-scm.org/D7668

status: make morestatus call out unresolved conflicts after update Differential Revision: https://phab.mercurial-scm.org/D7704

status: extract active-merge state for reuse Differential Revision: https://phab.mercurial-scm.org/D7667

fix: use rewriteutil.precheck() instead of reimplementing it Differential Revision: https://phab.mercurial-scm.org/D7687

split: use rewriteutil.precheck() instead of reimplementing it As you can see from the patch, I disagree with most of the comment saying that `rewriteutil.precheck()` is not worthwhile. Differential Revision: https://phab.mercurial-scm.org/D7686

rebase: use rewriteutil.precheck() instead of reimplementing it After this patch, there's still another place in `rebase.py`, in the `--stop` code path, that reimplements `rewriteutil.precheck()`. I couldn't fix that place because it `rewriteutil.precheck()` checks that there is only one dirstate parent, which fails because we have two parents at that point. I think it's incorrect that rebase leaves the user with two parents during conflicts, but changing that is way out of scope for this series. Differential Revision: https://phab.mercurial-scm.org/D7685