a very minor and probably should not happen in real-life since noone should check out git-annex branch (besides may be to fix up something manually) but I still do not see the point of scanning for unlocked/annexed files if init ran in that branch. ATM (8.20210428-270-g651fe3f39) I see

$> bash annex-scanning-git-annex
> set -eu
>> mktemp -d /home/yoh/.tmp/dl-XXXXXXX
> cd /home/yoh/.tmp/dl-yHMtPGl
> git clone --branch git-annex http://datasets.datalad.org/.git out
Cloning into 'out'...
Fetching objects: 11612, done.
> cd out
> git branch
* git-annex
> git annex init
init  (scanning for annexed files...)
ok
(recording state in git...)

real    0m0.622s
user    0m0.435s
sys 0m0.163s

so it reports scanning while in git-annex, although goes quickly (has only about 1000 keys there), so may be doesn't even actually do any scanning and it is just a matter of reporting?

Actually -- it made me think: is that scanning branch specific? then what would happen if e.g. master has no unlocked files in the tree and some other branch has unlocked files in the tree -- I could checkout/switch between branches without causing git-annex to redo its scanning and would require manual git annex init?

I feel this is unnecessary complexity and I've optimised the scans quite a lot in the meantime, so wontfix --Joey