Re: [PATCH v5 3/4] status: give more information during rebase -i
- From
Junio C Hamano <gitster@pobox.com>
- Date
- Jul 1, 2015, 16:36 UTC
- Message-ID
- <xmqqvbe3al87.fsf@gitster.dls.corp.google.com>
- In-Reply-To
- <CAPig+cSXzN84C4tOi5oZaK_1UJGcNXywE_-V7CyDQUMXox1Acw@mail.gmail.com>
Eric Sunshine <sunshine@sunshineco.com> writes:
Show 21 quoted lines
> I was about to mention the same shortcoming, but you beat me to it.
>
>> Perhaps be more strict and do this instead (without
>> leading strbuf_trim):
>>
>> if (!get_sha1_hex(split[1]->buf, sha1) &&
>> !strcmp(split[1]->buf + 40, " ") {
>> replace split[1] with "%s " abbrev
>> }
>
> Isn't it dangerous to assume that you can index 40 characters into
> split[1]? If (for some reason), the user botched the todo line such
> that the SHA1 is no longer a valid hex string, then split[1] will be
> that botched string, which might be shorter than 40 characters. For
> instance, if the user-edited todo line is:
>
> pick oops nothing
>
> then git-rebase--interactive.sh:transform_todo_ids() will leave "oops"
> as-is, since it can't unabbreviate it, and then this code will place
> "oops" into split[1].Yeah, that is why get_sha1_hex() is checked before we try to make sure buf[40] has " " in the code snippet you quoted.
Isn't that how && short-cut works?