Michael Hanke

Issue is not fixed by https://github.com/datalad/datalad/pull/4736

I can confirm that I also wasn't able to trigger it with git-annex outside datalad.

The following patch would turn 1min42s runtime into 270ms, and it would also remove the circular dependency, as well as the implied performance penalty https://github.com/datalad/datalad/issues/6941 ```diff diff --git a/datalad/support/gitrepo.py b/datalad/support/gitrepo.py...

I am in favor of starting this practice now. It boils to using the sphinx directives `versionadded`, `versionchanged`, and `deprecated` in the docstrings.

Related: - https://github.com/datalad/datalad/issues/5886 - https://github.com/datalad/datalad/issues/6509 I don't mind adding a feature to retry operations, a long as it does not impose substantial additional code complexity and that it is OFF...

In `-next` I use the following template ``` ### {{ cat }} - Describe change, possibly reference closed/related issue/PR. Fixes https://github.com/datalad/datalad-next/issues/XXX via https://github.com/datalad/datalad-next/pull/XXX (by @GITHUBHANDLE) ``` Any PR that needs...

Summary from devcall: It would be good to have a bot that reacts on the label `CHANGELOG-missing` being added to a PR, in this case add a simple changelog snippet...

When this is done, a patch similar to this one should be applied to the appveyor config (maybe there is something similar for travis): ```diff diff --git a/.appveyor.yml b/.appveyor.yml index...

Given that `siblings` already has: - `--annex-wanted` - `--annex-required` - `--annex-group` - `--annex-groupwanted` - `--no-annex-info` it does not seem too invasive to add `--annex-ignore` as an option directly.

I think we should take the core of the issue apart into the two pieces it seems to be comprised of. Let's assume for the following that `wtf` is synonymous...