Open Stack

On 1 January 2014 06:07, Joe Gordon <joe.gordon0 at gmail.com> wrote:
>
>

>
> I am not sure if this is the global .gitignore you are thinking of but this
> is the one I am in favor of:
>
> https://help.github.com/articles/ignoring-files#global-gitignore
>
>
> Maintaining .gitignore in 30+ repositories for a potentially infinite number
> of editors is very hard, and thankfully we have an easier way to do it.

This is a strawman argument: noone (that I know of) has proposed
adding all editors to all repositories. There are in reality a few
very common editors and having their extensions present in per
repository .gitignores does absolutely *no harm*. There is no reason
not to have sane and sensible defaults in our repositories.

If we are wasting time adding and removing patterns, then I think that
counts as a harm, so it is a sensible discussion to have to come to a
project standard, but the standard should be inclusive and useful, not
just useful for power users that have everything setup 'just so'. Many
contributors are using git for the first time when they contribute to
OpenStack, and getting git setup correctly is itself daunting [for new
users].

So I'm very much +1 on having tolerance for the top 5-10 editor
patterns in our .gitignores, -1 on *ever* having a bug open to change
this in any repository, and getting on with our actual task here of
writing fantastic code.

If folk *really* don't want editor files in .gitignore (and given the
complete lack of harm I would -really- like a explanation for this
mindset) then we could solve the problem more permanently: we know
what files need to be added - *.rst, *.py, *.ini, [!.]* and a few
others. Everything else is junk and shouldn't be added. By
whitelisting patterns we will support all editors except those whose
working file names match names we'd genuinely want to add.

-Rob

-- 
Robert Collins <rbtcollins at hp.com>
Distinguished Technologist
HP Converged Cloud