Re: [PATCH v3 3/7] builtin: patch-id: fix patch-id with binary diffs
- From
Junio C Hamano <gitster@pobox.com>
- Date
- Oct 17, 2022, 15:23 UTC
- Message-ID
- <xmqqk04y1nks.fsf@gitster.g>
- In-Reply-To
- <CAMKO5CvdDWEd6HPbkg7DP9bZMKNzcvmK+c1UPpuTk7vM1D8i9g@mail.gmail.com>
Jerry Zhang <jerry@skydio.com> writes:
Show 19 quoted lines
>> I thought I saw that a previous step touched diff.c to change how >> patch ID for a binary diff is computed to match what patch-id >> command computes? Now we also have to change patch-id? In the end >> output from both may match, but which one between diff and patch-id >> have we standardised on? > Er yeah let me see if I can simplify. > > Before: > Internal patch-id w/ unstable + binary was correct > Internal patch-id w/ stable + binary was broken > builtin patch-id w/ binary was broken > > After: > Internal patch-id w/ unstable + binary is correct > Internal patch-id w/ stable + binary is now correct > builtin patch-id w/ binary is now correct > > So the "standard" actually came from the one working case from > "before", which was the diff.c logic + unstable.
OK.
The question was meant to help you improve the log message, as it is something a future reader of "git log" would wonder after reading them. I think including something that makes it easy for readers to arrive at the summary above themselves by reading the log message would be a very much welcome change.
Thanks.