Michael Hanke

Here is another batch of constraints: https://github.com/datalad/datalad-gooey/pull/297 - `EnsureMapping` - `EnsureStr(regex)` - `EnsureGitRefName` - `EnsurePath` #6139

Here is another batch of constraints: https://github.com/datalad/datalad-gooey/pull/305 - `EnsureIterableOf` plus enhanced - `EnsureListOf` - `EnsureTupleOf` The latter two are only done because -core has less capable implementation, but otherwise the...

The target has shifted to datalad-next. With https://github.com/datalad/datalad-next/pull/116 there is now a consolidated RF/rewrite of the entire constraint system as a self-contained sub-package. It includes most of the structural improvements...

Not sure what is going on here: ```sh (datalad-dev) mih@meiner /tmp % cd top (datalad-dev) mih@meiner /tmp/top (git)-[master] % datalad create -d . sub1 add(ok): sub1 (dataset) add(ok): .gitmodules (file)...

The tuning is largely impossible to use, because it influences all clones -- we discussed that before. However, I was talking about compaction that happens while `save` is still running....

I am staring at another attempt to save 132k files (few kilobytes each). The progressbar suggests that this will take between a half and a full day. Transferring 3x the...

> However, for `save` it's probably more important to switch to `WitlessRunner` in the first place and only then have a new assessment of this aspect. Which translates to "stop...

Re git annex add FILES vs git annex add .: We have to keep in mind that passing input arguments as such is what we came from a long time...

FTR: https://github.com/datalad/datalad/pull/4850 closes the `_git_custom_command()` chapter.

I think we can close this. Another wave of performance updates will certainly come with updated estimates. At that point, this issues can serve as a "FOI" style anecdote.