[PATCH 2 of 4] profiling: Adding profiling.output config variable

[Matt Mackall]

On Thu, 2009-04-02 at 19:27 +0900, Nicolas Dumazet wrote:
> -_-' wrong mail.
> 
> The corrective patch is supposed to correct patch [1 of 4].
> 
> I'm being told by Dirkjan that I should
> 1) include doc within the same csets
> 2) probably add a few tests for the whole submission
> 
> 
> 
> On a more general patch submission policy, it is at least the second
> time I mail patches and {realize/I'm told} something is wrong.
> This generates unnecessary noise, I'm afraid. (sorry!)
> 
> I am thinking of less noisy ways for me to make my beginner mistakes.
> I guess that instead of trying to correct already-submitted set of
> patches, I should:
> 1) sit
> 2) modify properly the whole set
> 3) submit it to a public MQ
> 4) ask for review(s) on IRC
> 5) modify the MQ until patch set is satisfying
> 
> then, and only then, submit patches to the mailing-list?

Reposting patches is not a big deal, especially with small series or
single patches.

I'd really like to see version numbers on these things though so I can
tell which sets I can delete.

I'd also like to move away from [PATCH 0 of n] posts as they quite
consistently have important detail that ought to be in the changesets
themselves. 

-- 
http://selenic.com : development and support for Mercurial and Linux