Prevent false positives in initials as the are commonly used in names, e.g.
Margaret E. Hamilton - which obviously do not end sentences. Check whether a
period character `.' is preceded by at least two characters. This should save
us from false positives when linting.
* guix/lint.scm(check-description-style)[check-end-of-sentence-space] Add
condition.
* tests/lint.scm: Add test case.
Change-Id: I42a1365aaaed2afc7308b88ebd4b0720ad362761
Signed-off-by: Ludovic Courtès <ludo@gnu.org>
Fix linter warnings for the following:
- packages that belong to some programming language or ecosystem,
e.g. python-foo or texlive-bar,
- packages whose names end in a version distinction, e.g. wlroots-0.16 and
- packages where the software's real name contains an underscore `_'
character where our package name replaced that with a hyphen `-',
e.g. wpa_supplicant and wpa-supplicant-minimal.
* guix/lint.scm (check-description-style)[check-proper-start]: Add conditions.
* tests/lint.scm: New tests.
Change-Id: Ifc9f5cda04db59e460e287cd93afae89c7f17e3c
Signed-off-by: Ludovic Courtès <ludo@gnu.org>
Until now, ‘save-origin’ would be called only when given a
<git-reference>. With this change, ‘save-origin’ gets called for other
version control systems as well.
* guix/lint.scm (swh-response->warning): New procedure, formerly in
‘check-archival’.
(vcs-origin, save-package-source): New procedures.
(check-archival)[response->warning]: Remove.
Call ‘save-package-source’ in both the Git and the non-Git cases.
* tests/lint.scm ("archival: missing svn revision"): New test.
Change-Id: I535e4ec89488faf83bfa544d5e4935fa73ef54fb
While this method is new and nar-sha256 ExtIDs are currently available
only for new visits, it is fundamentally more reliable than the other
methods, which is why it comes first.
* guix/lint.scm (check-archival)[lookup-by-nar-hash]: New procedure.
Call ‘lookup-by-nar-hash’ before the other lookup methods.
* tests/lint.scm ("archival: content available")
("archival: content unavailable but disarchive available")
("archival: missing revision")
("archival: revision available"): Add a 404 response corresponding to
the ‘lookup-external-id’ request.
* tests/lint.scm ("archival: nar-sha256 extid available"): New test.
Change-Id: I4a81d6e022a3b72e6484726549d7fbae627f8e73
"CC=gcc" is almost always incorrect; people often just don't
notice the incorrectness because they are compiling natively.
For an exception, see tzdata.
"guix style" partially made things worse, so I partially ignored it.
* guix/lint.scm (check-compiler-for-target): New linter.
* tests/lint.scm
("compiler-for-target: unconditional CC=gcc is unacceptable")
("compiler-for-target: looks through G-expressions")
("compiler-for-target: (cc-for-target) is acceptable")
("compiler-for-target: CC=gcc is acceptable when target=#false"):
Test it.
Signed-off-by: Ludovic Courtès <ludo@gnu.org>
Suggested by Maxim Cournoyer <maxim.cournoyer@gmail.com>
and Simon Tournier <zimon.toutoune@gmail.com>.
* guix/lint.scm (check-archival): Add 'local-file?' clause. Clarify
message in case (package-source package) is not an origin.
* tests/lint.scm ("archival: not an origin"): New test.
emacs-build-system sets #:tests? #f by default, so the linter shouldn't warn
if #:tests? #t is set for packages using emacs-build-system. Likewise for
texlive-build-system.
* guix/lint.scm (check-tests-true): Do not warn if the build system
is emacs-build-system or texlive-build-system.
* tests/lint.scm
("tests-true: #:tests? #t acceptable for emacs packages")
("tests-true: #:tests? #t acceptable for texlive packages"): New tests.
Fixes: <https://issues.guix.gnu.org/50299>
Reported-by: Maxim Cournoyer <maxim.cournoyer@gmail.com>
Signed-off-by: Maxim Cournoyer <maxim.cournoyer@gmail.com>
With the switch to "ustar" format in commit
bdf5c16ac0, the maximum file length has
increased.
* guix/lint.scm (check-patch-file-names): Adjust margin used to check for
patch file lengths. Increase allowable patch file length appropriate to new
tar format. Extend warning to explain that long files may break 'make dist'.
* tests/lint.scm: Update tests accordingly.
This is a followup to e171182a20.
* tests/lint.scm ("description: invalid Texinfo markup")
("synopsis: valid Texinfo markup"): Add call to 'identity' to avoid
triggering a syntax error.
* guix/import/stackage.scm (<stackage-lts>, <snapshot>)
(<stackage-package>): New record types and JSON mappings.
(lts-info-packages, stackage-package-name)
(stackage-package-version): Remove.
(lts-package-version): Rename 'pkgs-info' to 'packages'; assume
'packages' is a list of <stackage-package>.
(stackage->guix-package): Use 'stackage-lts-packages' instead of
'lts-info-packages'. Rename 'packages-info' to 'packages'.
(latest-lts-release): Likewise.
(stackage-package?): Rename to...
(stackage-lts-package?): ... this. Adjust to new API.
(%stackage-updater)[pred]: Update accordingly.
* tests/lint.scm ("haskell-stackage"): Add "snapshot" entry in JSON
snippet.
This is a follow-up to commit 9c5e5ca1c0.
The previous package was called “ghc-x” which is not available on Stackage,
instead change it to “ghc-pandoc” which does exist, and adjust its version.
* tests/lint.scm ("haskell-stackage"): Add additional metadata for the
package; change package name to “ghc-pandoc”; and change to version to
“100.0”.
Reported-by: Tobias Geerinckx-Rice <me@tobias.gr>
Signed-off-by: Tobias Geerinckx-Rice <me@tobias.gr>
When using 'wrap-program', "bash" (or "bash-minimal") should be
in inputs. Otherwise, when cross-compiling, 'wrap-program' will use
a native bash instead of the cross bash and the 'patch-shebangs' won't
be able to correct this.
Tobias Geerinckx-Rice is added to the copyright lines because
a part of the "straw-viewer" package definition is included.
This linter detects 365 problematic package definitions at time
of writing.
* guix/lint.scm
(report-wrap-program-error): New procedure.
(check-wrapper-inputs): New linter.
(%local-checkers)[wrapper-inputs]: Add the new linter.
("explicit #:sh argument to 'wrap-program' is acceptable")
("'check-wrapper-inputs' detects 'wrap-program' without \"bash\" in inputs")
("'check-wrapper-inputs' detects 'wrap-qt-program' without \"bash\" in inputs")
("\"bash\" in 'inputs' satisfies 'check-wrapper-inputs'")
("\"bash-minimal\" in 'inputs' satisfies 'check-wrapper-inputs'")
("'cut' doesn't hide bad usages of 'wrap-program'")
("bogus phase specifications don't crash the linter"): New tests.
Signed-off-by: Mathieu Othacehe <othacehe@gnu.org>
There have been a few patches to the mailing list lately
not respecting this, and this linter detects 630 package
definitions that could be modified to support the --without-tests
package transformation.
* guix/lint.scm
(check-optional-tests): New linter.
(%local-checkers)[optional-tests]: Add it.
* tests/lint.scm
(package-with-phase-changes): New procedure.
("optional-tests: no check phase")
("optional-tests: check hase respects #:tests?")
("optional-tests: check phase ignores #:tests?")
("optional-tests: do not crash when #:phases is invalid")
("optional-tests: allow G-exps (no warning)")
("optional-tests: allow G-exps (warning)")
("optional-tests: complicated 'check' phase")
("optional-tests: 'check' phase is not first phase"): New tests.
Signed-off-by: Mathieu Othacehe <othacehe@gnu.org>
Note: this merge actually changes the 'curl' and 'python-attrs' derivations,
as part of solving caf4a7a277 and
12964df69a respectively.
4604d43c0e (gnu: gnutls@3.6.16: Fix cross-compilation.) was ignored because it
cannot currently be tested.
Conflicts:
gnu/local.mk
gnu/packages/aidc.scm
gnu/packages/boost.scm
gnu/packages/curl.scm
gnu/packages/nettle.scm
gnu/packages/networking.scm
gnu/packages/python-xyz.scm
gnu/packages/tls.scm
* guix/lint.scm (check-tests-true): New linter.
(%local-checkers)[tests-true]: Add it.
* tests/lint.scm ("tests-true: #:tests? must not be set to #t")
("tests-true: absent #:tests? is acceptable")
("tests-true: #:tests? #f is acceptable")
("tests-true: #:tests? #t acceptable when compiling natively"): Test it.
Signed-off-by: Mathieu Othacehe <othacehe@gnu.org>
* guix/lint.scm (lookup-disarchive-spec): New procedure.
(check-archival): When 'lookup-content' returns #f, call
'lookup-disarchive-spec'. Call 'lookup-directory' on the result of
'lookup-directory'.
* guix/download.scm (%disarchive-mirrors): Make public.
* tests/lint.scm ("archival: missing content"): Set
'%disarchive-mirrors'.
("archival: content unavailable but disarchive available"): New test.
As per section '16.4.2 Package Naming' in the manual, use hyphens
instead of underscores in package names.
* guix/lint.scm (check-name): Check whether the package name contains
underscores.
* tests/lint.scm ("name: use underscore in package name"): New test.
Signed-off-by: Ludovic Courtès <ludo@gnu.org>
Previously, test cases could fail if some process was listening
at a hard-coded port. This patch eliminates most of these potential
failures, by automatically assigning an unbound port. This should
allow for building multiple guix trees in parallel outside a build
container, though this is currently untested.
The test "home-page: Connection refused" in tests/lint.scm still
hardcodes port 9999, however.
* guix/tests/http.scm
(http-server-can-listen?): remove now unused procedure.
(%http-server-port): default to port 0, meaning the OS
will automatically choose a port.
(open-http-server-socket): remove the false statement claiming
this procedure is exported and also return the allocated port
number.
(%local-url): raise an error if the port is obviously unbound.
(call-with-http-server): set %http-server-port to the allocated
port while the thunk is called.
* tests/derivations.scm: adjust test cases to use automatically
assign a port. As there is no risk of a port conflict now,
do not make any tests conditional upon 'http-server-can-listen?'
anymore.
* tests/elpa.scm: likewise.
* tests/lint.scm: likewise, and add a TODO comment about a port
that is still hard-coded.
* tests/texlive.scm: likewise.
Signed-off-by: Ludovic Courtès <ludo@gnu.org>
In particular, "<origin> patches: same file name -> no warnings" would
have caught the issue which was fixed in commit
21887021b9.
* tests/lint.scm (patches: file names): Rename this test case...
("file patches: different file name -> warning"): ... to this.
("file patches: same file name -> no warnings")
("<origin> patches: different file name -> warning")
("<origin> patches: same file name -> no warnings"): New test cases.
Signed-off-by: Ludovic Courtès <ludo@gnu.org>
* guix/lint.scm (check-haskell-stackage): New procedure.
(%network-dependent-checkers): Add 'haskell-stackage' checker.
* guix/import/hackage.scm (%hackage-url): New variable.
(hackage-source-url, hackage-cabal-url): Use it in place of a
hard-coded string.
* guix/import/stackage.scm (%stackage-url): Make it a parameter.
(stackage-lts-info-fetch): Update accordingly.
* tests/lint.scm ("hackage-stackage"): New test.
Until now the 'source' checker would look at URL for 'url-fetch' origins
but not for 'git-fetch' origins.
* guix/lint.scm (check-source): Add case for 'git-reference?'.
* tests/lint.scm ("source, git-reference: 301 -> 200"): New test.
* guix/lint.scm (check-vulnerabilities): Add 'package-vulnerabilities'
optional parameter.
* tests/lint.scm ("cve: one vulnerability"): Use it instead of 'mock'.
* tests/cve.scm (vulnerability): Use 'make-struct/no-tail' instead of
'make-struct', which is deprecated.
* tests/lint.scm ("cve: one vulnerability")
("cve: one patched vulnerability")
("cve: known safe from vulnerability")
("cve: vulnerability fixed in replacement version")
("cve: patched vulnerability in replacement"): Likewise.