Re: [PATCH 01/16] midx-write: handle noop writes when converting incremental chains
- From
Taylor Blau <me@ttaylorr.com>
- Date
- Mar 31, 2026, 21:43 UTC
- Message-ID
- <acw/79TEXFs2cZyJ@nand.local>
- In-Reply-To
- <20260330223346.GA41843@coredump.intra.peff.net>
On Mon, Mar 30, 2026 at 06:33:46PM -0400, Jeff King wrote:
> Now we iterate through each layer individually. We assume that the total > count matches ctx->nr, otherwise the midx struct is totally corrupt. We > could count them and double-check, I guess, but I don't think there's > much point (it would be a BUG() if they were mismatched).
Yeah, I agree. The omitted context:
if (ctx->nr != midx->num_packs + midx->num_packs_in_base)
goto out;tells us that if we get to this spot, we expect our 'ctx->nr' to be equal to 'midx->num_packs + midx->num_packs_in_base'.
As soon as we find a non-already-MIDX'd pack, we 'goto out', so I don't think it makes much sense to build up a count and check it at the end even after we know we can't optimize out the write.
We could check that the ->num_packs of each of the tip's base MIDXs match ->num_packs_in_base, but the MIDX internals handle that for us by how we construct that field, since num_packs_in_base is *not* read from the on-disk MIDX.
Thanks, Taylor