Re: [RFC PATCH, WAS: "weird diff output?" 2/2] xdiff: implement empty line chunk heuristic

From: Jacob Keller <jacob.keller@gmail.com>
Date: Apr 15, 2016, 21:15 UTC
Message-ID: <CA+P7+xqe4ng9-gn1DRqhjebRXuHXbqZ6f3QsJ798k6DRo3bYyQ@mail.gmail.com>
In-Reply-To: <xmqqzisuaa3t.fsf@gitster.mtv.corp.google.com>

On Fri, Apr 15, 2016 at 1:30 PM, Junio C Hamano <gitster@pobox.com> wrote:

Show 26 quoted lines

> Jacob Keller <jacob.keller@gmail.com> writes:
>
>> On Fri, Apr 15, 2016 at 1:06 PM, Junio C Hamano <gitster@pobox.com> wrote:
>>
>>> I actually do not think these knobs should exist when the code is
>>> mature enough to be shipped to the end users.
>>>
>>> Use "diff.compactionHeuristics = <uint>" as an opaque set of bits to
>>> help the developers while they compare notes and reach consensus on
>>> a single tweak that they can agree on being good enough, and then
>>> remove that variable before the code hits 'next'.
>>>
>>> Thanks.
>>
>> I was under the impression that we would want a longer lived
>> configuration until we had enough data to say whether it was
>> helpful to make it default. I guess i had thought it would need to
>> be longer lived since there may be cases where it's not optimal
>> and being able to turn it off would be good?
>
> Once you start worrying about "some cases this may misbehave", a
> configuration variable is a wrong mechanism to do so anyway.  You
> would need to tie this to attributes, so the users can say "use this
> heuristics for my C code, but do not apply it for my AsciiDoc
> input", etc.
>

I think this makes perfect sense to apply this as an attribute, however.. why isn't the current diff algorithm done this way?

Thanks, Jake

> What you have is a pure developer support; aim to come up with "good
> enough" way, giving developers an easier way to experiment with, and
> remove it before the feature is shipped to the end user.