[PATCH 0 of 1 RFC] --- and scissor marks in mails inlining patches?

Matt Mackall mpm at selenic.com
Thu Apr 22 12:39:29 CDT 2010


On Thu, 2010-04-22 at 18:26 +0900, Nicolas Dumazet wrote:
> Hello!
> 
> This patch is related to issue2148, where an user complained that some git
> patches were wrongly parsed. For example, a mail using such header:
> 
> ----------cut-here-start-----------------------
> >From 955a3853069a0d7029668ecb27f9300ba6f84da0 Mon Sep 17 00:00:00 2001
> From: Denis Defreyne <denis.defreyne at stoneship.org>
> Date: Tue, 13 Apr 2010 11:49:08 +0200
> Subject: [PATCH 1/4] updated sass test to make it work with sass 3
> 
> ---
>  test/filters/test_sass.rb |    6 +++---
>  1 files changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/test/filters/test_sass.rb b/test/filters/test_sass.rb
> ----------cut-here-end-----------------------
> 
> We should be able to exclude the stats.
> 
> Following a discussion on IRC, it seems that we do not _always_ want to
> remove line after "---". Afterall, until then this line was considered legal
> and this might cause problem.
> Hence the idea of an option.

Can we try to actually look for a diffstat rather than just a cut line?
In other words, how strict can we easily make the test and still exclude
all of git's diffstat noise?

We need a default that does the right thing the vast majority of the
time, and failing that, does the safe thing. Otherwise, it will be of no
use: users will forget to use the option until it's too late.

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




More information about the Mercurial-devel mailing list