Re: git diff annoyance / feature request

[Boaz Harrosh <bharrosh@panasas.com>]

On 08/25/2011 01:27 PM, Junio C Hamano wrote:
> Boaz Harrosh <bharrosh@panasas.com> writes:
> 
>> I mean. The label "try_again" is not at all unique in my file. As a
>> reader I would like to see where is that code going to. The function
>> name is a unique file identifier that tells me exactly where the change
>> is going. The label is not. (It's not freaking BASIC)
>>
>> I bet all this was just inherited from diff. Would it be accepted if
>> I send a patch to fix it? What you guys think a goto label makes any
>> sense at all?
> 
> The default tries to mimic what GNU used to do when we added the feature.
> 
> The diff.*.xfuncname configuration variable is there exactly for people
> like you to tweak what we use for hunk headers. Please experiment with it
> and if you come up with a better set of patterns, people may want to copy
> it and use it themselves. we may even consider updating the built-in
> default with your patterns, once they got adopted by wider audiences.
> 

Thanks, I'll investigate it sounds very interesting.

> Personally, I would have to say that the source wouldn't be using too many
> labels with the same name for this behaviour to be problematic, especially
> if it is not freaking BASIC ;-), so...

The Linux Kernel is full of "goto out" or "goto err" its a common error handling
practice. I actually like it because it taps onto a known pattern.

Now the patch tell me @@@ lable out: !! that's not very useful I would say

Thanks I'm sure I can shape it up the way I like it
Boaz