For the record, docker tests pass with this version of swh.core, so it should be OK in terms of bw compat (should be...)

Jenkins job DCORE/gitlab-builds #247 succeeded in 56 sec.See Console Output, Blue Ocean and Coverage Report for more details.

I had to squash a number of revisions to prevent having a piece of the git history with failed tests (and it was hard to rework), so the main revis...

Jenkins job DCORE/gitlab-builds #246 succeeded in 56 sec.See Console Output, Blue Ocean and Coverage Report for more details.

Jenkins job DCORE/gitlab-builds #245 succeeded in 1 min 1 sec.See Console Output, Blue Ocean and Coverage Report for more details.

Jenkins job DCORE/gitlab-builds #244 succeeded in 1 min 8 sec.See Console Output, Blue Ocean and Coverage Report for more details.

So the last update do replace the ":" with a "." (with an explicit flag to specify the argument is a config path)....

we use this entry point to declare all the backends provided by a 'swh_package', and we use the api/logic that a backend which cls is postgresql is...

Jenkins job DCORE/gitlab-builds #243 succeeded in 57 sec.See Console Output, Blue Ocean and Coverage Report for more details.

to repeat my comment on swh-scrubber!74: could this be changed to always use the swh.core namespace? eg. f"swh.core.db.classes.{swh_package}"...

The config key being masking_db/blocking_db instead of db for the relevant proxies was an explicit attempt to avoid mixing them up in a nested con...

Using colons to represent paths in a nested dict configuration seems quite peculiar, most known syntaxes (e.g. elasticsearch, hiera, jq, ...) use ...

Using colons to represent paths in a nested dict configuration seems quite peculiar, most known syntaxes (e.g. elasticsearch, hiera, jq, ...) use d...

One thing I don't like with this is the need for the "config path" to target the db connection string entry (i.e. 'storage:steps:0:maksing_db') rat...

Jenkins job DCORE/gitlab-builds #242 succeeded in 1 min 8 sec.See Console Output, Blue Ocean and Coverage Report for more details.

Q: wouldn't it be better to store swh.storage:postgresql instead of storage:postgresql?...

there is room for code factorization in swh/core/cli/db.py...

Jenkins job DCORE/gitlab-builds #241 succeeded in 57 sec.See Console Output, Blue Ocean and Coverage Report for more details.

Jenkins job DCORE/gitlab-builds #240 succeeded in 56 sec.See Console Output, Blue Ocean and Coverage Report for more details.

Is it already implemented or is it for another diff?...