bug#22647: RFE: upgrade 'sed' RE's to include perlre syntax.

[Linda A. Walsh]

Hey anyone -- I know it's been a while though I don't wanna be 
pushy or such.. ;-)

But, um, anyone know where or what happened
to this change that may have gone in 1.5-2 years or so
ago?  It may simply be I don't have a late enough
version of sed. It's rpm says:

Source RPM  : sed-4.2.2-5.1.2.src.rpm
Build Date  : Thu 25 Sep 2014 01:15:55 AM PDT
Vendor      : openSUSE
Distribution: openSUSE 13.2

Thanks!
(probably, out of touch)
linda

p.s. - Had to work to try to make the quoting legible,
so original is in HTML, but am including txt as
well, but have no idea how clear it will be
considering how the HTML looks... ;-)


On 2014/07/24 at 19:29, Linda A. Walsh wrote:
> On 2014/07/24 at 09:39, Jose E. Marchesi wrote:
> > On 2014/07/22 at 07:44, Paolo Bonzini wrote:
> > > Il 22/07/2014 10:24, Linda A. Walsh ha scritto:
> > > >  Both simple atoms
> > > > such as "\d" for [0-9] or "\S" for non-space characters are supported
> > > > in awk and have been added to grep with the -P switch.
> > > >         
> > > > Would it be possible to add a -P type switch to 'sed' so that
> > > > PCRE-compat RE's could be used? 
> > >       
> > > Yes, indeed part of the work is already done and part of the GNU sed
> > > tarball (search for REG_PERL). Though the option was -R, not -P.
> > >       
> > > Paolo 
> >     
> > I want to release a new version of sed this August (it is time!) and I
> > would love to include this functionality on it. I think it is not too
> > late to use -P instead of -R, to keep some uniformity with the other
> > tools. 
>   
> I can try... :)
> Jose, in the meanwhile can you update gnulib in the sed repository?
> Paolo
>   
> > On 2014/07/24 at 09:52, Jose E. Marchesi wrote:
> > >  Done. 
> >   
>   
> Awesome. I don't have a heavy preference on -R v. -P, but trying to
> keep similar function switches the same between tools would be a bonus.
>   
> It is REALLY a headache remember what the switch is to specify the
> "delimiter" char in sort vs. cut, et al. I always seem to pick the
> wrong one.