Re: [PATCH v2 0/5] commit-graph: fsck zero/non-zero generation number fixes

From: Junio C Hamano <gitster@pobox.com>
Date: Aug 11, 2023, 19:28 UTC
Message-ID: <xmqqr0o9s7im.fsf@gitster.g>
In-Reply-To: <20230811175854.GA2816191@coredump.intra.peff.net>

Jeff King <peff@peff.net> writes:

Show 7 quoted lines

> This applies on top of yours, but probably would replace patches 2, 4,
> and 5 (the flip-flop case isn't even really worth testing after this,
> since the message can obviously only be shown once).
>
>  commit-graph.c          | 42 +++++++++--------------------------
>  t/t5318-commit-graph.sh | 18 ++-------------
>  2 files changed, 13 insertions(+), 47 deletions(-)

Quite an impressive amount of code reduction.  I obviously like it.

One very minor thing is that how much value are we getting by reporting the object names of one example from each camp, instead of just reporting a single bit "we have commits not counted and also counted their generations, which is an anomaly".

Obviously it does not matter. Even if we stopped doing so, the code would not become much simpler. We'd just use a word with two bits instead of two pointers to existing in-core objects, which does not have meaningful performance implications either way.

Thanks.