[GSoC 2010] The 1st week

[GSoC 2010] The 1st week

[Bo Yang]

Hi,

   This is the first week of gsoc 2010 coding phrase, and it is a good
time to summary the previous work and make a plan for the following
two weeks.

What's done:
1. Get familiar with Git code base and have read the code of 'git
blame', revision walk, 'git log'.
2. Try to make some patches for git:
    1. * by/blame-doc-m-c (2010-05-06) 1 commit, in 'next'
    2. * by/log-follow (2010-05-10) 4 commits, in 'next'
    3. the '--graph improvement', in discuss and I think ready for inclusion.
3. Start to write some code for the line level browser feature. And it
is in the 'playaround' branch in http://repo.or.cz/w/line.git . Now,
the code can parse multiple line ranges for multiple files in only one
revision.

What's following next 2 weeks:
In next 2 weeks, I will try to make up a version of line level browser
that supports code modification trace and this version will print all
the diff hunks that touch the interesting lines.

Regards!
Bo
-- 
My blog: http://blog.morebits.org

Re: [GSoC 2010] The 1st week

[Jeff King]

On Mon, May 24, 2010 at 10:57:15PM +0800, Bo Yang wrote:

> What's done:
> 1. Get familiar with Git code base and have read the code of 'git
> blame', revision walk, 'git log'.
> 2. Try to make some patches for git:
>     1. * by/blame-doc-m-c (2010-05-06) 1 commit, in 'next'
>     2. * by/log-follow (2010-05-10) 4 commits, in 'next'
>     3. the '--graph improvement', in discuss and I think ready for inclusion.
> 3. Start to write some code for the line level browser feature. And it
> is in the 'playaround' branch in http://repo.or.cz/w/line.git . Now,
> the code can parse multiple line ranges for multiple files in only one
> revision.

You didn't mention

  4. Contributing to other discussion of --follow on the list.

It is a small thing, perhaps, but I am very happy to see a GSoC student
actively participating in the community.

Speaking of which, I haven't finished polishing the --follow patches
from our earlier thread. I was taking a look at making history
simplification (including --parents) work. Are you going to be doing any
work on that this summer? If so, I will probably just drop it for now
and let you deal with it. :)

-Peff

Re: [GSoC 2010] The 1st week

[Bo Yang]

On Tue, May 25, 2010 at 2:24 PM, Jeff King <peff@peff.net> wrote:
> On Mon, May 24, 2010 at 10:57:15PM +0800, Bo Yang wrote:
>
>> What's done:
>> 1. Get familiar with Git code base and have read the code of 'git
>> blame', revision walk, 'git log'.
>> 2. Try to make some patches for git:
>>     1. * by/blame-doc-m-c (2010-05-06) 1 commit, in 'next'
>>     2. * by/log-follow (2010-05-10) 4 commits, in 'next'
>>     3. the '--graph improvement', in discuss and I think ready for inclusion.
>> 3. Start to write some code for the line level browser feature. And it
>> is in the 'playaround' branch in http://repo.or.cz/w/line.git . Now,
>> the code can parse multiple line ranges for multiple files in only one
>> revision.
>
> You didn't mention
>
>  4. Contributing to other discussion of --follow on the list.
>
> It is a small thing, perhaps, but I am very happy to see a GSoC student
> actively participating in the community.

Thanks! :)

> Speaking of which, I haven't finished polishing the --follow patches
> from our earlier thread. I was taking a look at making history
> simplification (including --parents) work. Are you going to be doing any
> work on that this summer? If so, I will probably just drop it for now
> and let you deal with it. :)

Yeah, the line level browser will touch the parents rewriting things
and I think the feature of line level history is a super-set of the
'--follow' things, so you can leave it with me. And in my plan, I will
complete this in next month. :)

Regards!
Bo
-- 
My blog: http://blog.morebits.org

Re: [GSoC 2010] The 1st week

[Jeff King]

On Tue, May 25, 2010 at 04:58:11PM +0800, Bo Yang wrote:

> Yeah, the line level browser will touch the parents rewriting things
> and I think the feature of line level history is a super-set of the
> '--follow' things, so you can leave it with me. And in my plan, I will
> complete this in next month. :)

Great, I will not worry about the parent rewriting problem, then. I will
probably polish and submit the multiple-file follow when I get a chance,
though.

-Peff

Re: [GSoC 2010] The 1st week

[Albert Krawczyk]

On 25/05/2010 4:24 PM, Jeff King wrote:
> Speaking of which, I haven't finished polishing the --follow patches
> from our earlier thread. I was taking a look at making history
> simplification (including --parents) work.
>    
I've been looking at getting --follow --parents to print correctly, 
however the Git code is way over my head. If somebody else smarter then 
me could have a look at that, it would be amazing...

I will keep trying, but I don't know how long it'll take me to get my 
head around the Git diff generation.

Albert