Scott Palmer wrote:
> On Jun 10, 2005, at 4:18 PM, Max Bowsher wrote:
>
>> Eric Gorr wrote:
>>
>>> Perhaps then it should be either an explicit command (svn
>>> addunversioned) or a flag on svn add.
>>
>> A new command, because it should be a combination of delete and add.
>
> Why not a flag or two to the existing 'svn add' and 'svn delete'?
>
> E.g.
> svn add --unversioned
> svn delete --missing
That UI works ok too, I guess... but it's 2 commands, when very often you
will want both behaviours simultaneously.
>>> and take --exclude patterns (see rsync for a good
>>> implementation of exclude patterns) so the user can avoid having
>>> files versioned that still should not be.
>>
>> Hmm. Really needed?
>
> It would help
>
>> Doesn't svn:ignore fulfil this purpose already?
>
> Not really. Since svn:ignore is a property set on the versioned
> folder it could easily not include things that shouldn't be versioned
> but aren't cumbersome enough that any specific user cared that they
> showed up in the status command. Also there is no control of what
> files someone might put in their WC, so the global (all users) nature
> of svn:ignore can't possibly know enough for the specific WC that the
> command is executed on. E.g. users might have little scripts that
> are specific to their WC.
Then they can sort out some standard naming pattern in order apply
svn:ignore to them.
Would they really want to remember to type --exclude foo *every* time?
If we did decide to go ahead with some kind of --exclude, it would make more
sense to do it generally, and make it available to *every* recursive
command, not just this one.
Max.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Sat Jun 11 00:07:06 2005