On Mon, Mar 15, 2010 at 19:35:33 +0000, Alexey Levan wrote:
> Actually, there's no problem with the fixSubPaths function - it just
> filters out incorrect absolute paths, and it does it just fine. However,
> the real problem is how this function is (mis)used.

Thanks!  I really appreciate it when Darcs hackers take the extra time
out to look at the wider context of their work, which seems to be what
you have done (I hadn't even considered the users of fixSubPaths; was
just thinking of sticking a fail in there)

> There's no consistent policy of using fixSubPaths, and if you use darcs
> command with invalid absolute paths, the following may happen:
>   * everything is OK, because invalid paths are handled explicitly
> (Record.lhs, Rollback.lhs)
>   * everything is OK, because command with no args isn't special case
> (Diff.lhs)
>   * you get message like "Nothing specified, nothing to do". Message is
> misleading, but nothing bad happens (Remove.lhs, Add.lhs)
>   * the worst: you get error (Move.lhs, try 'darcs move /foo /bar /baz'!).

I have not really checked the specifics of the code, but at a high
level, that sounds like the right sort of thing to do: harmonise,
modularise

-- 
Eric Kow <http://www.nltg.brighton.ac.uk/home/Eric.Kow>
PGP Key ID: 08AC04F9

Attachment: pgpKroGaAdU2n.pgp
Description: PGP signature

_______________________________________________
darcs-users mailing list
[email protected]
http://lists.osuosl.org/mailman/listinfo/darcs-users

Reply via email to