2008-03-15 22:22:35

by Artem S. Tashkinov

[permalink] [raw]
Subject: GIT commit description flag(s)

It would be nice if [kernel] git commits had a flag [a char] (or a
combination of) describing their main purpose which could be the following:

bug Fix [F]
Speed up/Performance [S]
Revert [R]
New functionality [N]

etc.

So a particular commit at
http://git.kernel.org/?p=linux%2Fkernel%2Fgit%2Ftorvalds%2Flinux-2.6.git;a=summary

will look like:

19 hours ago Ingo Molnar sched: fix fair sleepers [F] commit |
commitdiff | tree | snapshot

This way it will be much easier to keep track of important changes
happening to the kernel.


--

Thanks for your attention,

Artem S. Tashkinov


2008-03-16 10:11:39

by Johannes Weiner

[permalink] [raw]
Subject: Re: GIT commit description flag(s)

Hi,

"Artem S. Tashkinov" <[email protected]> writes:

> It would be nice if [kernel] git commits had a flag [a char] (or a
> combination of) describing their main purpose which could be the
> following:
>
> bug Fix [F]
> Speed up/Performance [S]
> Revert [R]
> New functionality [N]

> This way it will be much easier to keep track of important changes
> happening to the kernel.

You claim there are unimportant ones? :)

Hannes

2008-03-19 21:10:26

by Jan Engelhardt

[permalink] [raw]
Subject: Re: GIT commit description flag(s)


On Mar 16 2008 02:35, Artem S. Tashkinov wrote:
> It would be nice if [kernel] git commits had a flag [a char] (or a combination
> of) describing their main purpose which could be the following:
>
> bug Fix [F]
> Speed up/Performance [S]
> Revert [R]
> New functionality [N]

Usually you already give the type implicitly in the commit:

dfd347f... HID: fix comment in hid_input_report()
fa331ff... [S390] sclp_vt220: speed up console output for interactive work
d9452e9... [NETPOLL]: Revert two bogus cleanups that broke netconsole.
8727e28... m68k{,nommu}: Wire up new timerfd syscalls
b863ceb... [NET]: Add macvlan driver

So all is well.

2008-03-19 22:48:18

by Artem S. Tashkinov

[permalink] [raw]
Subject: Re: GIT commit description flag(s)

I quite agree with you that the description is already there, but such
flags could help easily review or grep commits in order to identify your
areas of interest.

Even your example has ambiguity: "Wire up new timerfd syscalls" - It's
not easy to understand whether this particular commit is a bugfix, or it
speeds up things, or the old code which has been cleaned up, etc.

Jan Engelhardt wrote:
>
> On Mar 16 2008 02:35, Artem S. Tashkinov wrote:
>> It would be nice if [kernel] git commits had a flag [a char] (or a
>> combination of) describing their main purpose which could be the
>> following:
>>
>> bug Fix [F]
>> Speed up/Performance [S]
>> Revert [R]
>> New functionality [N]
>
> Usually you already give the type implicitly in the commit:
>
> dfd347f... HID: fix comment in hid_input_report()
> fa331ff... [S390] sclp_vt220: speed up console output for interactive work
> d9452e9... [NETPOLL]: Revert two bogus cleanups that broke netconsole.
> 8727e28... m68k{,nommu}: Wire up new timerfd syscalls
> b863ceb... [NET]: Add macvlan driver
>
> So all is well.
>

2008-03-20 11:07:09

by Stefan Richter

[permalink] [raw]
Subject: Re: GIT commit description flag(s)

Artem S. Tashkinov wrote:
> I quite agree with you that the description is already there, but such
> flags could help easily review or grep commits in order to identify your
> areas of interest.
>
> Even your example has ambiguity: "Wire up new timerfd syscalls" - It's
> not easy to understand whether this particular commit is a bugfix, or it
> speeds up things, or the old code which has been cleaned up, etc.

The fundamental problems
- people sometimes chose bad summaries,
- we can't always predict the full impact of the patch,
- what's important for me isn't for you
cannot be solved by your proposed flags.

They can only be solved by
- reviewing patches WRT good changelogs,
- writing up extra summary release notes for areas of interest
after the fact.

>>> It would be nice if [kernel] git commits had a flag [a char] (or a
>>> combination of) describing their main purpose which could be the
>>> following:
>>>
>>> bug Fix [F]
>>> Speed up/Performance [S]
>>> Revert [R]
>>> New functionality [N]

New bug [B]
Attempted bug fix which doesn't really work, a.k.a. nice Try [T]
Performance regression, a.k.a Downgrade [D]
--
Stefan Richter
-=====-==--- --== =-=--
http://arcgraph.de/sr/