- 17 Jan, 2019 9 commits
-
-
Chris Lamb authored
-
Chris Lamb authored
Gbp-Dch: ignore
-
Chris Lamb authored
Gbp-Dch: ignore
-
Chris Lamb authored
-
Chris Lamb authored
-
Chris Lamb authored
Gbp-Dch: ignore
-
Chris Lamb authored
Gbp-Dch: ignore
-
Chris Lamb authored
-
The test runner terminated with exit code 0 when a tag listed in the 'onlyrun' option could not be found. This issue is mentioned in #919531, but does not close the bug. Signed-off-by:
Chris Lamb <lamby@debian.org>
-
- 16 Jan, 2019 5 commits
-
-
Chris Lamb authored
Add a check for empty executable files in PATH; these files do not do anything and produce no error message. (Closes: #919458)
-
As a way to help gather data and aid in future migration to /usr, as well as a means of detecting packages that installed files to / to help mount /usr that no longer need to, add an experimental pedantic-only check for files in / (/bin, /sbin, /lib*) that are not compatibility symlinks to the corresponding file in /usr. Note that this is a pedantic-only check, with no plans to migrate it to a warning or error in the near future and should not be used as the basis for any form of mass bug filing. Signed-off-by:
Chris Lamb <lamby@debian.org> -
Chris Lamb authored
-
Chris Lamb authored
-
Chris Lamb authored
Prune Uploaders field to the members of the "lintian" LDAP group with thanks to all who have contributed in the past. (We do respect Niels explicit removal in 15118c1c)
-
- 13 Jan, 2019 4 commits
-
-
Previously we used the extracted version of 'dplint' under autopkgtests. We correct the path so that the installed version is used in compliance with autopkgtest requirements. Signed-off-by:
Chris Lamb <lamby@debian.org> -
Improves on the previous commit if none of the paths in File::Find::Rule exist. In that case, the call to stat could fail. Substitutes 'time' instead. This would also triage #919096 but set HARNESS_EPOCH to a value more recent than necessary. Signed-off-by:
Chris Lamb <lamby@debian.org> -
The harness files are not installed into /usr/share/lintian so they were not found at testing time. Use the untarred testing code (but the installed version of Lintian). Signed-off-by:
Chris Lamb <lamby@debian.org> -
Paul Wise authored
-
- 12 Jan, 2019 2 commits
-
-
Chris Lamb authored
Split the wrong-path-for-interpreter into "wrong" and "incorrect" variants, the latter with a reduced severity and capturing incorrect (but working) shebangs such as "#!/usr/bin/env perl". Thanks to Ivo De Decker for the report. (Closes: #917790)
-
Chris Lamb authored
Gbp-Dch: ignore
-
- 11 Jan, 2019 10 commits
-
-
Chris Lamb authored
Update Debian Policy regex and manual references since this document moved to Sphinx. (Closes: #918963)
-
Chris Lamb authored
-
Chris Lamb authored
Prevent false-positives in maintainer-script-should-not-use-dpkg-maintscript-helper by adding optional support for ignoring sections automatically added by debhelper. (Closes: #917567)
-
Chris Lamb authored
Add a note that using a maintainer-script-should-not- prefix ensures that the tag will automatically pass the internal coverage tests.
-
Use TAP::Harness to run all Lintian tests. This unifies all tests. It has the advantage that all results appear at the bottom. It is therefore unlikely to miss a failure in the code quality scripts or in the internal harness tests. Previously, those results tended to be obscured by long output from the package tests. The short output shown when building via 'debuild' (presumably for lack of a tty) is no longer available. There may be a way to customize TAP::Harness, if that is important. Also, the command line setting $keep_going does not have any effect currently. Somewhat surprisingly, such a "bail-on-fail" setting is not available in Perl's standard TAP::Harness. It only requires a trivial subclass available on CPAN, but was not used for now to avoid adding another dependency. This can be addressed, if necessary. As for speed, running all tests through TAP::Harness is currently slower than the old runner. The difference is about 8 percent. This comes as a surprise since previous tests showed them about the same. Perplexingly, the parallel test preparation seems responsible. It is the only recent feature. Other more modern harness implementations such as Test2::Harness offer pre-loaded persistent runners. They could bring a speed gain that should be investigated in time. Signed-off-by:
Chris Lamb <lamby@debian.org> -
Test evaluated with Perl's 'prove' or with TAP::Harness must communicate their results and other status information using the "Test Anything Protocol" (TAP). TAP is a text-based protocol that works differently from the way information is conveyed by the existing persistent runners in Lintian, which use IO::Async::Channel. Converts the test modules to emit TAP messages using Test::More. Signed-off-by:
Chris Lamb <lamby@debian.org> -
Deletes old test scripts and installs the current generic TAP script into test working directories. The scripts can be evaluated using 'prove' or TAP::Harness. Signed-off-by:
Chris Lamb <lamby@debian.org> -
A generic test scripts that can be used for Lintian with TAP::Harness is available at t/runners/generic.t. The default setting allows the test preparation routines to find it. Then it is copied into each test's work directory before it is run. Signed-off-by:
Chris Lamb <lamby@debian.org> -
Adds a generic test script that can be installed in test working directories. It is then invoked via 'prove' or TAP::Harness. Requires that modules reporting results switch from a properly initialized Test::State to the "Test Anything Protocol", which is available from various modules such as Test::More. Signed-off-by:
Chris Lamb <lamby@debian.org> -
Chris Lamb authored
-
- 10 Jan, 2019 10 commits
-
-
Paul Wise authored
-
Chris Lamb authored
-
Chris Lamb authored
-
Chris Lamb authored
-
Chris Lamb authored
-
A recent commit series to split the test preparation from the test runner causes the following error: not a reference at /usr/share/perl5/IO/Async/Channel.pm line 197. The problem was likely added in commit 6ffd137a . This now passes (and also resolves) an explicit reference to the path that identifies each test. Signed-off-by:
Chris Lamb <lamby@debian.org> -
Chris Lamb authored
-
Chris Lamb authored
Apply a patch series from Felix Lechner to split the preparation of tests from the test runner itself. (MR: !122)
-
Simplifies the logic for process messaging in Test::State and elsewhere. The key is to issue multiple commands in the worker threads when changing a test's status. That may not be atomic, but the most unpleasant consequence should be messages out of sequence. None were observed so far. Signed-off-by:
Chris Lamb <lamby@debian.org> -
Harnesses are generally better at executing tests than at preparing test material. Due to the complexity of many Lintian tests, the two stages are better split. TAP::Harness may take over the running of the tests. The test preparation, on the other hand, will continue to run in parallel using IO::Async::Routing and a setup very much like the existing test runner. Now it just has its own event loop. Signed-off-by:
Chris Lamb <lamby@debian.org>
-