Re: [PATCH] pathspec: rename per-item field has_wildcard to use_wildcard

[Nguyen Thai Ngoc Duy <pclouds@gmail.com>]

On Thu, Apr 7, 2011 at 6:09 PM, Michael J Gruber
<git@drmicha.warpmail.net> wrote:
> Junio, we're in rc phase ;)

He was excited. It's good, we may get this feature sooner :)

>>  (2) Colon, open parenthesis, a comma separated list of words to name
>>      magic, close parenthesis, followed by the path, e.g. these are the
>>      long-hand counterparts to the examples in (1)
>>
>>      - ":(top)hello.c"
>>      - ":(top,noglob)hello.c"
>>      - ":(noglob,top)hello.c"
>>      - ":(noglob,top):hello.c"
>>      - ":()hello.c"
>>
>
> Do we need the parentheses? I guess we need them to have the magic start
> with a non-alpha, otherwise we could do with ":top,noglob:hello.c"
> (which I like but breaks those with a file named ":top:hello.c").

That has something to do with "easy to parse and catch error", I
think. At the moment ":hello/:world" is interpreted as "hello/:world".
Without parentheses, we need to look further for colon and may
accidentally misinterpret it.

> What about these:
>
> :/(noglob)hello.c
> :(top)!hello.c
>
> I.e. do we allow combinations?

I think we should. I don't like typing long pathspecs, just because
only remember mnemonic for "top" and happen to need "noglob" too.

> It seems you accept to break people with files named ":(top)hello.c" but
> not those with files named ":top:hello.c". Which is OK if it's a
> conscious decision, of course.

At least once we agree on colon being the magic introducer and general
rules, we can do these earlier in 1.7.x

 - warn people that ":hello.c" may change semantics in future
 - recommend to escape ("::hello.c" instead of ":hello.c")
-- 
Duy