git-p4: planned view wildcard support

git-p4: planned view wildcard support

[Matthew Mendell]

Hello,

Is there any plan for the git-p4.py script to support the "*" wildcard
in a p4 client?

The specific test that exposes this unsupported feature:
./t9809-git-p4-client-view.sh
...
ok 4 - unsupported view wildcard *

Thanks,
-Matt

Re: git-p4: planned view wildcard support

[Pete Wyckoff]

matthew.mendell@gmail.com wrote on Sun, 06 May 2012 16:46 -0700:
> Is there any plan for the git-p4.py script to support the "*" wildcard
> in a p4 client?
> 
> The specific test that exposes this unsupported feature:
> ./t9809-git-p4-client-view.sh
> ...
> ok 4 - unsupported view wildcard *

It hasn't been on anybody's priority list so far that I know, but
this would be not too hard to implement in theory.

The code changes would need to alter just one class.  In
View.Path, functions Path.find_wildcards, match_wildcards and
fill_in_wildcards, if you're interested in giving it a try.

Can you constrain your use of "*" a bit?  What makes these
wildcards difficult is all the edge cases.  For instance, is
there ever more than one "*"?  Is it combined with "..."?  Are
there multiple lines in the client spec with wildcards?  Some
abstracted examples would be interesting.

		-- Pete

Re: git-p4: planned view wildcard support

[Matthew Mendell]

On Mon, May 7, 2012 at 3:44 AM, Pete Wyckoff <pw@padd.com> wrote:
> matthew.mendell@gmail.com wrote on Sun, 06 May 2012 16:46 -0700:
>> Is there any plan for the git-p4.py script to support the "*" wildcard
>> in a p4 client?
>>
>> The specific test that exposes this unsupported feature:
>> ./t9809-git-p4-client-view.sh
>> ...
>> ok 4 - unsupported view wildcard *
>
> It hasn't been on anybody's priority list so far that I know, but
> this would be not too hard to implement in theory.
>
> The code changes would need to alter just one class.  In
> View.Path, functions Path.find_wildcards, match_wildcards and
> fill_in_wildcards, if you're interested in giving it a try.
>
> Can you constrain your use of "*" a bit?  What makes these
> wildcards difficult is all the edge cases.  For instance, is
> there ever more than one "*"?  Is it combined with "..."?  Are
> there multiple lines in the client spec with wildcards?  Some
> abstracted examples would be interesting.
>
>                -- Pete

Hi Pete,

Yeah, I have been perusing the code and would like to give it a shot.
Is it ok to initially add partial support for the easy cases?
My usage of "*" is straight forward:

//depot/project/* //my_clientspec/depot/project/*

Thanks,
-Matt

Re: git-p4: planned view wildcard support

[Pete Wyckoff]

matthew.mendell@gmail.com wrote on Mon, 07 May 2012 08:19 -0700:
> On Mon, May 7, 2012 at 3:44 AM, Pete Wyckoff <pw@padd.com> wrote:
> > matthew.mendell@gmail.com wrote on Sun, 06 May 2012 16:46 -0700:
> >> Is there any plan for the git-p4.py script to support the "*" wildcard
> >> in a p4 client?
> >>
> >> The specific test that exposes this unsupported feature:
> >> ./t9809-git-p4-client-view.sh
> >> ...
> >> ok 4 - unsupported view wildcard *
> >
> > It hasn't been on anybody's priority list so far that I know, but
> > this would be not too hard to implement in theory.
> >
> > The code changes would need to alter just one class.  In
> > View.Path, functions Path.find_wildcards, match_wildcards and
> > fill_in_wildcards, if you're interested in giving it a try.
> >
> > Can you constrain your use of "*" a bit?  What makes these
> > wildcards difficult is all the edge cases.  For instance, is
> > there ever more than one "*"?  Is it combined with "..."?  Are
> > there multiple lines in the client spec with wildcards?  Some
> > abstracted examples would be interesting.
> 
> Yeah, I have been perusing the code and would like to give it a shot.
> Is it ok to initially add partial support for the easy cases?
> My usage of "*" is straight forward:
> 
> //depot/project/* //my_clientspec/depot/project/*

That one does seem like it wouldn't be too hard.  Similar to
trailing "...", but matching only one path component.

As long as git-p4 can reliably detect when it can deal
with a *, I think partial support would be just fine.

I'll help think about test cases and such once you have
some code that's ready to look at.  Thanks!

		-- Pete