On Tue, Apr 30, 2013 at 10:28:21AM -0400, Marc Branchaud wrote:
> On 13-04-29 05:15 PM, Junio C Hamano wrote:
> > Marc Branchaud <marcnarc@xxxxxxxxxxx> writes:
> >
> >> This started out as an attempt to make the backward compatibility notes
> >> more parsable, but then I just kept going...
> >
> > Thanks.
> >
> >> * "git bundle" did not like a bundle created using a commit without
> >> - any message as its one of the prerequistes.
> >> + any message, as it is one of the prerequistes.
> >
> > This is actually saying a different thing.
> >
> > When you create a bundle, you can say "you can only unbundle this in
> > a repository that has commit X", with "git bundle create $name ^X Y
> > Z". Such a commit X is called the bundle's prerequisite. You can
> > have more than one prerequisite, e.g. "... ^X ^W Y Z".
> >
> > But if you create a bundle by using a commit that does not have any
> > message as X (i.e. the bundle's prerequisite), the "git bundle" did
> > not like to read the resulting bundle output.
> >
> > So <a commit <without any message> as its (bundle's) one of the
> > prerequisites> is what the original wanted to say. The rewrite
> > makes it read like "For a commit, having a message is a requirement
> > to be used in a bundle", at least to me.
>
> Thanks, I did get that wrong.
>
> CC'ing Lukas, who wrote the relevant commit (5446e33f35).
>
> How about:
>
> * "git bundle" can create a bundle that has a commit without a message as
> a prerequisite, but it could not work with such a bundle.
Looks fine to me.
Junio's version with the last part changed to "as one of its
prerequisites" sounds also good to me (in both cases, however, note the
missing "i" in "prerequisites").
A third suggestion:
"git bundle" erroneously bailed out when parsing a valid bundle
containing a prerequisite commit without a commit message.
>
> M.
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html