[darcs-users] darcs patch: Ignore junk more conservatively. (and 1 more)

Jason Dagit dagit at codersbase.com
Wed Oct 8 21:01:37 UTC 2008


On Wed, Oct 8, 2008 at 11:38 AM, David Roundy <daveroundy at gmail.com> wrote:

> On Wed, Oct 8, 2008 at 1:22 PM, Eric Kow <kowey at darcs.net> wrote:
> > On Wed, Oct 08, 2008 at 13:10:06 -0400, David Roundy wrote:
> >> What I don't know (and maybe it's hidden somewhere in those emails
> >> that I only skimmed) is what problem there is.  Why is it considered
> >> important for users to be able to write "Ignore-this: foobar" in their
> >> patches and have darcs display this to people who pull from them?
> >
> > I think it's just Zooko trying to be on the safe side.  He thinks it's
> > conceivable that users would actually put Ignore-this: into their patch
> > logs for daily use (I think it's a stretch, but who knows).
> >
> > The second thing he is trying to do is to make us more future-proof by
> > suggesting a way to do things so that we can add more metadata in the
> > future without confusing darcs 2.1.0
>
> I don't see that we've yet added any metadata, and don't see any
> reason to do so.  I think this is a red herring, and we should wait
> until there's a feature that would benefit from this.  Or at least we
> should ask whether there is some sort of hypothetical feature that it
> would benefit.  In the absence of any of those, I'd say we've got what
> we want.


Actually, people propose various per-patch features from time to time.
Allowing user-defined attributes on patches (metadata) would be a nice
pre-factor that allows us to address many of these concerns.  Take for
instance the proposal to have local-only patches.  These patches could have
metedata like:
X-Local-Only: True

Note: I'm using the metadata syntax compatible with Trent Buck's proposal.
It somehow missed Eric's summary, but the idea is to use X-Foo headers like
email uses.  Trent quoted the RFC but I'm too lazy to look it up.

As far as the concern about users putting Ignore-this: into their
> patch information, I don't see it as a compelling reason to add
> complexity to the code, especially since I consider hidden comments to
> be a very reasonable feature to add to darcs.


I think in the Jason/Trent proposal, it would be more like "X-Ignore-this:
....", and darcs could choose to ignore it.  Which attributes to ignore
could be defined in a prefs file.  There are lots of options to consider, I
think.

One final note, I would prefer if the user-defined attributes had their own
place in the patch format, but I seem to be the only person who feels
strongly about it.  So, I guess putting it into the comment field is OK for
now.

Jason
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.osuosl.org/pipermail/darcs-users/attachments/20081008/fda27441/attachment.htm 


More information about the darcs-users mailing list