Release v2.7.0pre1
#6872
Release v2.7.0pre1
#6872
Conversation
Codecov Report❌ Patch coverage is Additional details and impacted files@@ Coverage Diff @@
## support/2.7.0pre1 #6872 +/- ##
=====================================================
- Coverage 78.59% 77.47% -1.12%
=====================================================
Files 567 576 +9
Lines 43092 45707 +2615
=====================================================
+ Hits 33866 35409 +1543
- Misses 9226 10298 +1072 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
| ) | ||
| __license__ = 'MIT license, see LICENSE.txt file.' | ||
| __version__ = '2.6.3.post0' | ||
| __version__ = '2.7.0pre1' |
There was a problem hiding this comment.
Should this be?
| __version__ = '2.7.0pre1' | |
| __version__ = '2.7.0rc1' |
There was a problem hiding this comment.
So PEP does not anymore support pre (In old it still was). When released it is automatically converted to rc0 last time but I started this usage preX so I thought for these preleases I just stay consistent and use preX. No strong feelings about either way.
There was a problem hiding this comment.
I think it is very confusing to have a version of pypi different from __version__.
There was a problem hiding this comment.
Yes but it is also confusing to have a history (on git) of
2.7.0pre0
2.7.0rc1
so what do you prefer? Next time I want to start with rc0
There was a problem hiding this comment.
Well hopefully not many people are installing these versions so ultimately it doesn't matter I guess. :-) no strong opinions, with preference to switch to rc1 now.
Squashed commit at 2025-05-09 21:53 PR aiidateam#6793 introduced the cancelation of earlier kill actions. This had the problem if two kill commands are set in a sequence, the second kill action will cancel the first one which triggered the cancelation of the scheduler job within an EBM. The second kill command however did not retrigger the cancelation of the scheduler job. This bug appeared because we have two places where the killing logic is placed. More information about this can be found in PR aiidateam#6868 that fixes this properly refactoring the kill action. This PR only serves as a fast temporary fix with workarounds. Before this PR, when the killing command failed through the EBM, the scheduler job could not be cancelled through a kill anymore. Since we have now force-kill option to bypass the EBM, we can reschedule the cancelation of the scheduler job to gracefully kill a process.
Squashed commit at 2025-05-09 21:54 Add config pydantic model Add detect.py Add group-node-mapping Add dump logger Add dump engine Add dump managers Add facades Add utils Add changes to CLI Add changes to init, disable mypy for feature for now Add changes to docs Add changes to and additional tests [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci Fix bug in explicitly-groupd sub-workflows being filtered out for profile/group dumping Fix group validation exception on `verdi profile dump -G` and creation of empty dirs for deselected groups.
agoscinski
force-pushed
the
support/2.7.0pre1
branch
from
8055b5d to
3b96a2f
Compare
…iidateam#6867) Squashed commit at 2025-05-09 22:04 defer_build in Entity.Model class defer_build in Entity.Model pydantic class
Cherry-picked at 2025-05-09 22:06 This aims to avoid reconstruction of fields from parent models during MRO walks.
agoscinski
force-pushed
the
support/2.7.0pre1
branch
from
3b96a2f to
e4205e9
Compare
4 participants
In addition to the current state on main b618cb7 I included for production testing the PRs
verdi process killfix Regular killing reschedules a cancel of scheduler job #6870verdi {profile|group|process} dumpProfile data dumping #6723Please test again with this prerelease:
verdi process killcould bypass EBM by sequentially killing