Re: MIDX woes, was Re: [ANNOUNCE] Git v2.54.0-rc2
- From
Junio C Hamano <gitster@pobox.com>
- Date
- Apr 16, 2026, 20:12 UTC
- Message-ID
- <xmqqqzoezmt9.fsf@gitster.g>
- In-Reply-To
- <20260416194922.GA1887222@coredump.intra.peff.net>
Jeff King <peff@peff.net> writes:
> Yeah, I agree that re-wording would be a good idea here.
I actually no longer think so.
Show 9 quoted lines
> I think this > can happen post-release. We'd like for this to have happened long ago, > but it didn't. Doing it in 2.54 versus 2.55 does not really matter, as: > > 1. We still have to worry about 2.53 and earlier anyway, so one > version is not much in the grand scheme of things. > > 2. For midx v2, 2.54 knows how to read it anyway, so is not affected > by the problem. ;)
I agree 100% with this.
Looking at the later part of load_multi_pack_index_one() function, it is full of die() calls, not prepared to degrade gracefully at all.
The reason why I do not think reword is worth doing is that as far as this function (i.e., "Here is an MIDX file---please read it and give me struct multi_pack_index back") is concerned, these error returns are giving diagnosis on the nature of errors. Sometimes we cannot learn the length of the file, sometimes we find that the file is unreasonably short, sometimes it is not sorted properly, etc. None of the existing "error + cleanup" says "we are continuing anyway" and that is good. The decision to keep going without using midx file (or not) belongs to the caller, so if we wanted to say that, we should say so in the caller, not here.