RF: just pip install datalad-installer#5676
Conversation
|
Just commenting, have no checked yet. It would be good to know if a diff to the extension setup exists, and if so why, and if all extensions need to be modified. |
|
It is at least missing the ability to switch the installer version without editing .appveyor.yml. The feature was introduced in the extension setup the last time the installer behavior broke, and all extensions had to be fix before any CI would work again. install:
...
# deploy the datalad installer, override version via DATALAD_INSTALLER_VERSION
- cmd:
IF DEFINED DATALAD_INSTALLER_VERSION (
python -m pip install "datalad-installer%DATALAD_INSTALLER_VERSION%"
) ELSE (
python -m pip install datalad-installer
)
- sh: python -m pip install datalad-installer${DATALAD_INSTALLER_VERSION:-}
it would be great if the setups converge. |
the only not applied change is removal of datalad.plugin testing -- still in maint
…-annex installation way
it would be great if the setup didn't diverge ;-) copied |
Codecov Report
@@ Coverage Diff @@
## maint #5676 +/- ##
===========================================
- Coverage 90.32% 29.49% -60.83%
===========================================
Files 299 296 -3
Lines 42349 42313 -36
===========================================
- Hits 38251 12480 -25771
- Misses 4098 29833 +25735
Continue to review full report at Codecov.
|
|
hm, codecov offset is too big. #5688 against maint has only 0.1% . looking at diff in coverage - might relate to travis runs for high log level to not provide results, but also might relate to added |
|
ok, had to restart one appveyor on mac. now all green and codecov divergence is small. so we should be good to go |
mih
left a comment
There was a problem hiding this comment.
Wonderful, thx @yarikoptic for the updates!
master branch has its own ideas and binds to some older version
needs yet to be addressed, decided to just go with maint for now and simplify (not versioned depends etc)