RE: [PATCH] clarify documentation for remote helpers
- From
- David Turner <david.turner@twosigma.com>
- Date
- Aug 30, 2019, 17:45 UTC
- Message-ID
- <c6c737c6a37c4cf7aa77e5c5ffd7a86f@exmbdft7.ad.twosigma.com>
- In-Reply-To
- <xmqqv9uey2tb.fsf@gitster-ct.c.googlers.com>
Show 22 quoted lines
> -----Original Message----- > From: Junio C Hamano <gitster@pobox.com> > Sent: Friday, August 30, 2019 1:35 PM > To: David Turner <David.Turner@twosigma.com> > Cc: Martin Ågren <martin.agren@gmail.com>; Git Mailing List > <git@vger.kernel.org> > Subject: Re: [PATCH] clarify documentation for remote helpers > > David Turner <David.Turner@twosigma.com> writes: > > > I was confused, because I read "a file under GIT_DIR/objects/pack" > > to mean "just the filename". Some of the things that deal with packs > > take just the filename (e.g. --keep-pack for git repack). > > I'll fix the under under and add $, but I do want to clarify that it's > > the full path. > > I think that the phrase wanted to say that the file named with the option must > live under that directory, without any implication that the directory is used as > the base when a relative path is used. If the helper MUST give a full pathname > and a pathname relative to that directory is not accepted, it is a good idea to > spell it out (also if it must end with ".keep", that also should be documented--- > is there any other restrictions?).
The only other restriction I see is: in order for the connectivity-skipping optimization to be used, the file with s/.keep/.idx/, and the corresponding pack, must exist. Do you think that's worth mentioning? It seems to be implied by the rest of the text.