The "new" auto hinter relies on partial ordering to determine, when
what can migrate (and what needs to migrate at the same time). At the
same time, it leverages on "_compute_groups" to allow it to include
"removals" in its hints.
Signed-off-by: Niels Thykier <niels@thykier.net>
Avoid smooth-updating libraries in hints, when all of their reverse
dependencies will certainly disappear in the same hint.
Note that in "hint"-hint, reverse dependencies removed in the
following "full run" will not cause the smooth-updated library to be
removed. Instead these will still be removed in the end as usual, but
in some cases that is too late.
Signed-off-by: Niels Thykier <niels@thykier.net>
Rename find_upgraded_binaries into _compute_groups. The new method
will also compute what binaries will be updated in or added to testing
after migration.
Signed-off-by: Niels Thykier <niels@thykier.net>
- Added docstrings
- Replaced format by %
- Import ADT_EXCUSES_LABELS from autopkgtest.py instead of redefining it
- Renamed test to avoid name conflict with autopkgtest.py from britney
- Fixed some formatting
- A test for a package/version can be triggered by several distincts causes with different results. Modify the data structure to store this instead of only storing the last cause and result, leading to missing test results in excuses.
- Check if failures are regressions (there is at least one pass and migration is blocked) or if test has always been failing. In the latter case, migration is not blocked.
- Add colours to excuses.html to distinguishes the test results
- Point jenkins URL to lastBuild
In excuse_unsat_deps(), mark unsatisfiable dependencies not just in the HTML,
but also in the invalid_deps list. If we have any of those in
should_upgrade_src(), block the package and don't run the autopkgtest. This
avoid running tests for known-uninstallable packages, which just leads to
guaranteed failures, manual intervention of re-running tests after it becomes
installable, and spamming maintainers with a FAIL/PASS notification.
This reproduces the failure to hold back gccgo-4.9 due to the broken libgcc1
package. This only seems to happen when a new source package takes over an
existing binary package.
Colin Watson
Niels Thykier
Adam D. Barratt
Julien Cristau
Jean-Baptiste Lallement
Martin Pitt
Philipp Kern