[PATCH] changegroup: do not prompt to merge after adding only closed heads (issue2697)

[Adrian Buehlmann]

On 2011-04-14 23:08, Matt Mackall wrote:
> On Thu, 2011-04-14 at 22:40 +0200, Adrian Buehlmann wrote:
>> On 2011-04-14 22:20, Matt Mackall wrote:
>>> On Thu, 2011-04-14 at 22:01 +0200, Adrian Buehlmann wrote:
>>>> On 2011-04-14 21:09, Kevin Gessner wrote:
>>>>> When adding a changegroup adds N heads, M of which are closed, tell
>>>>> the user that "(+N heads, M closed)" were added to the repository.
>>>>
>>>> "When adding a changegroup adds N heads, .."
>>>>
>>>> has still the same grammar bug, I'd say.
>>>
>>> Parses here. "When [action] adds N heads".
>>
>> Ok. But at least it's hard to read.
>>
>> We shouldn't write law text.
>>
>>> But I also should remark that grammar/spelling issues in commit message
>>> bodies are way down on my list of priorities.
>>
>> I personally have made the observation that if there's even an error or
>> a bad description for the change, then the probability of a bug/flaw in
>> the actual code change is indeed higher.
>>
>> I can dig out actual evidence for that, if you like.
>>
>> Call it "change message smell". It works quite well as an indicator.
> 
> On the other hand, we don't really want to require all our contributors
> to have perfect English. Some of them are Americans, after all.
> 

My English is horrible, that's why I probably had a problem with the
text in the first place. I'm a native Swiss (born in Berne, thus the
slow thinking), as some may already know.

Change message text is more often read than written, much the same like
code. So I'd like to make it easy to read, even for people who's primary
language is not English.

And, for what it's worth, I hate those often seen nebulous three or four
word sentence change messages in complicated patches (which is not the
case here). It's like people have to hide something.

If even the guru can't explain his change in simple text, then who can
review it?