Junio C Hamano <gits...@pobox.com> writes:

>> +    for (i = 0; i < revs.nr; i++) {
>> +            if (bad_seen)
>> +                    string_list_append(&states, terms->term_good.buf);
>> +            else {
>> +                    bad_seen = 1;
>> +                    string_list_append(&states, terms->term_bad.buf);
>> +            }
>> +    }
> This is certainly different from the original.  We used to do one
> "bisect_write" per element in revs in the loop; we no longer do that
> and instead collect them in states list.  Each element in these two
> lists, i.e. revs.item[i] and states.item[i], corresponds to each
> other.
> That explains why the variable is renamed to states.  We haven't
> seen enough to say if this behaviour change is a good idea or not.

Ahh, I misread the original.  It accumulates the writes to be
executed in $eval and does so at the end.  So there is no change in

So please ignore that point in the previous message.  That leaves
only the following points:

 * Perhaps 'retval' with 'goto finish' pattern?

 * ref_exists()?  Perhaps use skip_prefix(head, "refs/heads/, &branch)?

 * if (clean-state) { return -1 }?

 * Is comment on trap relevant here?

Sorry, and thanks.
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Reply via email to