Open yarikoptic opened 2 years ago
Given that @yarikoptic said he
would like to continue prior discussion in #5517 (comment) to get a final decision before we cut a release with new
foreach-dataset
command
I wanted to comment to get this needed discussion going.
I don't have any strong opinions, though. I can see some advantages in a dedicated --cmd
parameter, if only for making the command call more verbose/readable.
I have no strong opinion given pros
cons
stated in the description. May be leaning more toward "just let it be positional" due to the pros, and the main motivation for making it --cmd
(consistency on having path
as the first (or all) positional(s)) being a bit superficial since nobody raised even desire in having path
restriction so far.
I would like to continue prior discussion in https://github.com/datalad/datalad/pull/5517#issuecomment-966651403 to get a final decision before we cut a release with new
foreach-dataset
command.That would open the possibility to add
path
as the positional argument, since itpath
is used to limit the scope of an operationcons'ish like
git submodule foreach
takes it as a positional argumentforeach-dataset
is notrun
, so most likely it is ok to have them a bit differentan "alternative" was voiced in https://github.com/datalad/datalad/pull/5517#issuecomment-966651403 -- to add
--path
kwarg in case we do decide to add path-based limiting. But it still sounds to me suboptimal that would make it inconsistent with the rest of the interfaces.WDYT?