| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
| |
Once again they broke the API, this time around the .captures() call.
They changed it by a lot, returning a dict of capture names to captures.
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
| |
Resolves: https://github.com/pkgcore/pkgcheck/issues/695
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
| |
Resolves: https://github.com/pkgcore/pkgcheck/issues/693
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
| |
Resolves: https://github.com/pkgcore/pkgcheck/issues/659
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
| |
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
| |
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
|
|
|
| |
Now that most ebuilds use PEP-517 mode, we can relax the check a little
to handle cases where the variable is set in the eclass (for example
tree-sitter based ebuilds).
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
| |
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
| |
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
|
| |
Fixes: 4a7a02f60c9f87ad85670e0e798c5aadb8096c92
Resolves: https://github.com/pkgcore/pkgcheck/issues/679
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
|
|
|
|
| |
It might sometimes not find any intersection of eclasses EAPIs, in which
it is better to not report anything then to report a false positive or
worse, to explode with exception.
Resolves: https://github.com/pkgcore/pkgcheck/issues/679
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
e.g.
pkgcheck scan --net -c HomepageUrlCheck --timeout 0 app-admin/chrootuid
at checkout 127160ac611d39cc6bb2ca21fcf99a086fe2b176
Python's ftplib raises a ValueError with timeout=0. Use timeout=None to
put the underlying socket in blocking mode. See
https://docs.python.org/3/library/socket.html#socket.socket.settimeout
for legal timeout values.
Signed-off-by: Lucio Sauer <watermanpaint@posteo.net>
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
|
|
| |
We can use the new tree-sitter-bash python package, which has the
library, so we can simplify the build a lot.
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
|
| |
Resolves: https://github.com/pkgcore/pkgcheck/issues/671
Resolves: https://github.com/pkgcore/pkgcheck/issues/672
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
|
|
|
| |
Catch cases where new ebuilds are committed with old EAPIs. This is
checked during `--commits` stage.
Resolves: https://github.com/pkgcore/pkgcheck/issues/666
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
| |
Resolves: https://github.com/pkgcore/pkgcheck/issues/668
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
| |
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This reverts commit 9103513e26f9f2aeade5b563a49697c0e2665e3e.
I originally added --no-find-copies in e688357bdc5773009bb2e106075d9852f2513f89
to suppress git being too clever which ends up confusing our git integration, as
we're not really interested in copies. I was going to use --no-find-copies-harder
but I figured --no-find-copies sounded better as it was less specific and for
our purposes here, we don't want copies at all, so why not?
But --no-find-copies isn't a boolean option, it's just that until git commit
5825268db1058516d05be03d6a8d8d55eea5a943 ('parse-options: fully disable option abbreviation with PARSE_OPT_KEEP_UNKNOWN'),
it tolerated it and expanded it to --no-find-copies-harder. Oops!
Let's use --no-find-copies-harder as we originally should have, as the only
available option to control the behavior, modulo setting a similarity % threshold
with --find-copies=n.
We're fine to do this as, quoting the git-log docs:
"[...] and options applicable to the git-diff[1] command to control how the changes each commit introduces are shown."
Bug: https://bugs.gentoo.org/924718
Bug: https://github.com/pkgcore/pkgcheck/issues/663
Signed-off-by: Sam James <sam@gentoo.org>
Closes: https://github.com/pkgcore/pkgcheck/pull/664
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
|
|
|
|
| |
It broke with git-2.43.2, until further upstream work by Sam James, I'll
just revert it.
Bug: https://bugs.gentoo.org/924718
Reverts: e688357bdc5773009bb2e106075d9852f2513f89
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
|
|
|
|
| |
Most notably, this includes NotOpenSSLWarning:
https://github.com/urllib3/urllib3/issues/3020#issuecomment-1785873825
Signed-off-by: Anna “CyberTailor” <cyber@sysrq.in>
Closes: https://github.com/pkgcore/pkgcheck/pull/661
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Gentoo developers are rejecting routine version bumps for ebuild
variables being defined in a different order than skel.ebuild.
This new lint ensures pkgcheck identifies these problems before
we waste developer time.
Regarding tests, in spite of the massive diff, all that's been done
is re-ordering the variables to avoid introducing new style warnings
into existing tests.
Signed-off-by: Anthony Ryan <anthonyryan1@gmail.com>
Closes: https://github.com/pkgcore/pkgcheck/pull/645
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
| |
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
| |
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
pkgcore's `category_dirs` is a fossil that shouldn't be
relied upon. My recollection is this property was from
before `profiles/categories` existed, but it's just
a listdir() intersection against known directories.
PMS mandates categories be in `profiles/categories`, thus
that functionality breaks compatibility.
It will be removed from pkgcore in future semver breaks
for that reason.
Signed-off-by: Brian Harring <ferringb@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
| |
This is beyond corner case, but if someone has a non directory
in place of where a category is supposed to be, pkgcore and portage
will silently suppress this.
Pkgcore will be modified to emit a warning for this, but in the meantime,
it's a cheap check, thus add it.
Signed-off-by: Brian Harring <ferringb@gmail.com>
|
|
|
|
|
| |
Resolves: https://github.com/pkgcore/pkgcheck/issues/650
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
|
|
|
|
| |
Catch cases where we depend on pkgmoved package, and instead of showing
NonexistentDeps and NonsolvableDepsIn* errors, show DependencyMoved
result, with nice and simple suggestion on action to take.
Resolves: https://github.com/pkgcore/pkgcheck/issues/649
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
|
|
| |
Catches multiple arguments passed to function, and colon separated path.
Resolves: https://github.com/pkgcore/pkgcheck/issues/644
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Check for cases where a dependency uses conditional use dependency
with unknown USE flag. PMS states:
It is an error for an ebuild to use a conditional use dependency when
that ebuild does not have the flag in IUSE_EFFECTIVE.
Bug: https://bugs.gentoo.org/921841
Requested-by: Sam James <sam@gentoo.org>
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
| |
Resolves: https://github.com/pkgcore/pkgcheck/issues/642
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
| |
Requested-by: Sam James <sam@gentoo.org>
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
|
|
|
|
| |
* Notably handles the particular case with gitlab package hosting.
Bug: https://github.com/pkgcore/pkgcheck/issues/636
Signed-off-by: Alfred Wingate <parona@protonmail.com>
Closes: https://github.com/pkgcore/pkgcheck/pull/637
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
|
|
|
|
|
| |
* Not setting it triggers a DeprecationWarning in python3.12.
* Despite tarballs by git-archive being trusted, there isn't benefit in
trusting it as the tarballs don't use metadata that would require it.
Signed-off-by: Alfred Wingate <parona@protonmail.com>
Closes: https://github.com/pkgcore/pkgcheck/pull/638
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
|
|
|
|
|
| |
Came up as example from elisp eclass, which has automagic calling of
functions from readme.gentoo-r1 eclass if they are found in env. While
weird, this is legal so let's add a simple whitelist support for this,
so it would be simple to extend in future.
Reported-by: Ulrich Müller <ulm@gentoo.org>
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
| |
Resolves: https://github.com/pkgcore/pkgcheck/issues/633
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
| |
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
| |
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
|
|
| |
Add some commands which should never be called in ebuild to the banned
list.
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Upon initial failure to get the git diff-tree output (so not the hot
flow, but sad failure flow), try to catch output of incorrectly
configured git remote (the remote should have a correct HEAD configured).
Recommend the user to run `git remote set-head origin -a` to solve it.
Resolves: https://github.com/pkgcore/pkgcheck/issues/608
Resolves: https://github.com/pkgcore/pkgdev/issues/107
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
| |
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
|
|
| |
Resolves: https://github.com/pkgcore/pkgcheck/issues/613
Resolves: https://github.com/pkgcore/pkgcheck/issues/612
Resolves: https://github.com/pkgcore/pkgcheck/issues/545
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
|
|
|
| |
Catch ``has_version --host-root`` and ``best_version --host-root``
calls, which are not allowed in EAPI>=7.
Resolves: https://github.com/pkgcore/pkgcheck/issues/630
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
|
| |
Closes: https://bugs.gentoo.org/365691
Resolves: https://github.com/pkgcore/pkgcheck/issues/624
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
| |
Resolves: https://github.com/pkgcore/pkgcheck/issues/622
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
|
| |
Resolves: https://github.com/pkgcore/pkgcheck/issues/625
Closes: https://bugs.gentoo.org/596616
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
| |
Resolves: https://github.com/pkgcore/pkgcheck/issues/626
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
| |
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
|
| |
Signed-off-by: Thomas Bracht Laumann Jespersen <t@laumann.xyz>
Closes: https://github.com/pkgcore/pkgcheck/pull/620
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
| |
Signed-off-by: Thomas Bracht Laumann Jespersen <t@laumann.xyz>
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
git log
I currently have a local git hack to allow configuring git to default to
--find-copies-harder because it's *extremely* useful when working on ebuild
repositories (prompted by a discussion with Eli Schwartz).
Unfortunately, this can confuse pkgcheck's git intergration because it'll
call `git log` like:
```
git log --name-status --diff-filter=ARMD -z --pretty=tformat:%n%h%n%ct cc5b3b9f134a070c548faa4e3de17d615497d0b3..origin/HEAD
```
and get nothing back because (I think) git is interpreting some changes as copies
rather than renames or new files.
Explicitly pass options to disable finding copies, even though normally this
isn't necessary, to keep things working.
Signed-off-by: Sam James <sam@gentoo.org>
Closes: https://github.com/pkgcore/pkgcheck/pull/618
Signed-off-by: Arthur Zamarin <arthurzam@gentoo.org>
|