Re: GSoC Project | Improvise git bisect

From: Christian Couder <christian.couder@gmail.com>
Date: Mar 21, 2016, 17:53 UTC
Message-ID: <CAP8UFD1_mxej3YE_vU7DaXEsDpu-VA1YZtG5zkhLE2NXHpfL+Q@mail.gmail.com>
In-Reply-To: <CAFZEwPNd3X9JocrrJa-d4mV-U_MZdMEq=gMxByZ9iQd6Tz4w2Q@mail.gmail.com>

Hi Pranit,

On Mon, Mar 21, 2016 at 8:29 AM, Pranit Bauva <pranit.bauva@gmail.com> wrote:

Show 18 quoted lines

> On Mon, Mar 21, 2016 at 12:48 PM, Johannes Schindelin
> <Johannes.Schindelin@gmx.de> wrote:
>> Hi Pranit,
>>
>> On Sun, 20 Mar 2016, Pranit Bauva wrote:
>>
>>> I could first move individual functions to bisect--helper.c.
>>
>> My suggestion would be to give it a try already with some functionality
>> you deem small enough to move to the bisect--helper within a day or so. It
>> is always good to test the waters like that, and to include this early
>> work in the proposal, also to assess (and to let the reviewers assess) how
>> feasible the project is.
>
> Sure! I will start with check_term_format().
>
> Also, I think implementing a new algorithm for bisect along with this
> would be too big for a GSoC project

Yeah, I also think so.

> and I am no algorithm expert. I
> will stick to --first-parent along with incremental rewrite.

My opinion is that an incremental rewrite by itself is big enough for a GSoC. And it might be difficult to do something else that is bisect related at the same time. So if you want to do the rewrite, just focus on it. If you ever have some time left we will easily find other interesting bisect related improvements.

If you really want to do "--first-parent" then you should probably add "Improve git bisect terms" to it in your proposal.

Best, Christian.