move PGSU out to separate package #3892
Conversation
|
/update-requirements |
|
Starting 'update-requirements' workflow for this branch. |
ltalirz
force-pushed
the
issue_2837_extract_postgres
branch
from
cde1a78 to
361a513
Compare
ltalirz
force-pushed
the
issue_2837_extract_postgres
branch
from
361a513 to
66b0ea1
Compare
One of the main issues with `verdi quicksetup` is that it is non-trivial to figure out how to connect as the PostgreSQL superuser in a wide variety of operating systems and PostgreSQL setups. This PR factors out the code which takes care of this connection into a separate package, which can then be tested on a wide variety of setups using CI.
ltalirz
force-pushed
the
issue_2837_extract_postgres
branch
from
66b0ea1 to
ffbe6a2
Compare
|
/update-requirements |
|
Starting 'update-requirements' workflow for this branch. |
|
@csadorf It seems the update-requirements workflow already fails at the "git checkout" stage... see https://github.com/aiidateam/aiida-core/actions/runs/71254296 |
Codecov Report
@@ Coverage Diff @@
## develop #3892 +/- ##
=========================================
Coverage 78.00% 78.00%
=========================================
Files 457 457
Lines 33830 33708 -122
=========================================
- Hits 26390 26295 -95
+ Misses 7440 7413 -27
Continue to review full report at Codecov.
|
|
@sphuber This is ready for review |
I had the same issue for #3599. |
| echo.echo_info('database {} already exists!'.format(dbname)) | ||
| if not click.confirm('Use it (make sure it is not used by another profile)?'): | ||
| dbname = click.prompt('new name', type=str, default=dbname) |
There was a problem hiding this comment.
I realize now (and not sure if this was already the case) that the code here is actually tightly coupled to the CLI. Is this code only used from the CLI, or should it be? Maybe not for this PR, but it would be good to think about this and see if we can make this code more general or really just move it to aiida.cmdline if it really is just meant for the CLI
There was a problem hiding this comment.
Yeah, there are many pieces like this in the code, where certain functions can be called in "interactive" mode and then start prompting.
Once we have a clear path forward how to deal with those, we should go through the entire codebase once and move them.
fixes #2837
fixes #2912
fixes #2588
fixes #3879
One of the main issues with
verdi quicksetupis that it is non-trivialto figure out how to connect as the PostgreSQL superuser in a wide
variety of operating systems and PostgreSQL setups.
This PR factors out the code which takes care of this connection into a
separate
pgsupackage, which can then be tested on a wide variety of setupsusing CI.