Yaroslav Halchenko

"Fusion" approach to changelog generation

9

Had to look how bids-specification deals with [src/CHANGES.md](https://github.com/bids-standard/bids-specification/blob/HEAD/src/CHANGES.md) generation: coded in https://github.com/bids-standard/bids-specification/blob/HEAD/.circleci/config.yml#L176 . In a nutshell critical difference is that workflow produces/commits/pushes changelog entry as soon as PR is merged....

-c: we do not allow for empty value or spaces in it?

2

while reviewing https://github.com/datalad/datalad/pull/6864/files#diff-e49e10eb8fa981f878166eae3246a7d39816638f95f2f7dd2c9186aee63829b5R258 spotted that regex for config setting is ``` re.compile(r'[^\s]+\.[^\s]+=[\S]+' ``` so I wondered - why `\S` in `[]` -- no need, right? - we are not allowing...

boost minimal supported git-annex from 8.20200309

2

just now, while filing datalad/datalad-ria#49 I realized that we are still supporting > 2 yo git-annex. Well, we kinda support it since we do have some number of test skips...

in cron jobs happens only in a single matrix run ``` $> datalad foreach-dataset "git grep 'FAILED .*test_nested_pushclone_cycle_allplatforms' | grep cron || :" foreach-dataset(ok): /mnt/datasets/datalad/ci/logs/2022 (dataset) 03/cron/20220603T193647/173dc6f/travis-14100-failed/14.txt:FAILED ../core/distributed/tests/test_push.py::test_nested_pushclone_cycle_allplatforms foreach-dataset(ok): /mnt/datasets/datalad/ci/logs/2022/01...

spotted that recent cron job failed, looked at https://app.travis-ci.com/github/datalad/datalad/jobs/576842693 and confirmed historically: ``` $> datalad foreach-dataset "git grep 'FAILED ../datalad/distributed/tests/test_ria_basics.py::test_version_check' | grep cron || :" foreach-dataset(ok): /mnt/datasets/datalad/ci/logs/2022 (dataset) 03/cron/20220603T193647/173dc6f/travis-14100-failed/11.txt:FAILED ../datalad/distributed/tests/test_ria_basics.py::test_version_check...

test_basic_scenario_local_url -- FAILs on cron run on master

1

in `_DL_ANNEX_INSTALL_SCENARIO="miniconda --batch git-annex=8.20200309 -m conda"` matrix run -- so likely just annex incompatibility Spotted in a failed run on master https://app.travis-ci.com/github/datalad/datalad/jobs/563044933 ``` ====================================================================== FAIL: datalad.customremotes.tests.test_datalad.test_basic_scenario_local_url ---------------------------------------------------------------------- Traceback (most recent...

CRON run https://app.travis-ci.com/github/datalad/datalad/builds/253266640 https://app.travis-ci.com/github/datalad/datalad/jobs/576842687 with - cmd:annex: 10.20220525-gf1fdc90 - cmd:bundled-git: UNKNOWN - cmd:git: 2.37.1 ``` ../datalad/core/distributed/tests/test_clone.py::test_clone_into_dataset No output has been received in the last 10m0s, this potentially indicates a stalled...

git bash on windows is not detectable as "interactive"

2

See https://stackoverflow.com/questions/40912072/git-for-windows-mintty-sys-stdout-isatty-returns-false for explanations etc. I do not see any clean workaround ``` DataLad@DESKTOP-1FTT289 MINGW64 /c/tmp/datasets.datalad.org (adjusted/master(unlocked)) $ python -c "import sys; print(sys.stdin.isatty())" False ```

test_command_fail_1 fails on debian testing py 3.9 (only)

2

``` _____________________________ test_command_fail_1 ______________________________ def test_command_fail_1(): # Expect that the return code of a failing command is caught, # that None is returned as result. bc = BatchedCommand( cmd=py2cmd( """...

prevent submodules which start with '-'

6

just noted `git status` complaining on one of shub subdatasets. It seems (didn't check in detail) that we are ok with it (since iirc we do parsing of .gitmodules ourselves)...