bug-cvs
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: CVS Keyword expansion patch


From: Scott James Remnant
Subject: Re: CVS Keyword expansion patch
Date: 12 Mar 2003 17:26:24 +0000

On Wed, 2003-03-12 at 17:14, Mark D. Baushke wrote:

> Scott James Remnant <address@hidden> writes:
> 
> > On Wed, 2003-03-12 at 15:12, Mark D. Baushke wrote:
> > 
> > I think you misunderstand the intent of the patch, it only looks at the
> > colon immediately after the keyword.
> 
> Yes, I did misunderstand the patch (I guess I must needed a bit more
> caffeine or something to kickstart my brain this morning... :-)
> 
Morning?!  It's at least lunchtime before my brain ever boots :-)

> For the case at hand, I am under the impression that PERL allows you to
> use ${Header::fields}{$foo} as an alternative method of specifying
> $Header::fields{$foo} in most perl 5.001 and beyond releases.
> 
> I suspect it be safer to code defensively rather than worry which
> version of cvs a user might be using with that code.
> 
Aye, but it's something even the most experienced Perl users trip
over...

> I think I understand the intent of the patch and I have no problems with
> the general goal of making rcs keyword expansion do what users mean it
> to do. In this case, that appears to mean restricting the patterns that
> are considered for keyword substitution from $keyword$ and $keyword:.*$
> to something other criteria like '$keyword$' '$keyword:$' and 
> '$keyword: .*$' right?
> 
'$keyword(:[^:]?.*)?$' is probably the best way to describe it :)

> So a more general patch might be that the character after a colon must
> be either whitespace or a $ in order to allow the keyword expansion to
> take place.
> 
I was worried that this was getting a bit too close to breaking possible
code, and thought the conservative "No ::" was a bit better.

> Note: I have only recently been given the ability to commit to the
> repository, so I'd still like to have Larry or Derek bless all features
> that cause a behavior change like this one before they get committed.
> 
> If you want to get them to bless your patch, you will want to provide a
> sanity.sh test case for the new feature as well as the changes needed to
> the documentation files.
> 
*nods* I thought I'd wait until I got the impression whether it was a
yay or nay idea before going ahead and doing the rest ;)

Scott
-- 
sjr  software and systems engineer
     demon internet, thus plc




reply via email to

[Prev in Thread] Current Thread [Next in Thread]