Re: Less greedy pattern match

"Ed Morton" <morton@[EMAIL PROTECTED]
> wrote in message 
news:480B46EC.9080607@[EMAIL PROTECTED]
>
>
> On 4/19/2008 11:15 AM, Cesar Rabak wrote:
>> Ed Morton escreveu:
>>
>>>On 4/18/2008 3:06 PM, pk wrote:
>>><snip>
>>>
>>>>After a bit of tweaking, it seems that the two cases can actually be 
>>>>folded
>>>>into a single body:
>>>>
>>>>{for(i=m+1;i<=NF;i+=2){printf "%s",$i};
>>>> if (NF) {m=(NF%2==m)?1:0}
>>>> if (m==0) print""}
>>>>
>>>>Not elegant as the gawk solution, but at least it should be standard.
>>>>
>>>
>>>At least you got it in one script! Best I could come up with was a
2-pass
>>>approach to turn all the searched-for patterns into a single char 
>>>(control-B in
>>>this case for no particular reason) first and then use that as the RS:
>>>
>>>awk '{$1=$1}1' FS='</?t1>' OFS=^B file | awk 'NR%2' RS=^B ORS=
>>>
>>
>> Notwithstanding, a very elegant solution as it is runnable as one
liner!
>
> This has been a bit of an eye opener for me to the annoyance of 2 
> inconstencies
> in the standard 
> (http://www.opengroup.org/onlinepubs/009695399/utilities/awk.html):
>
> 1) You can have an RE as the FS but not the RS
> 2) You can negate a character in an RE but not a sub-expression in an RE
>
> Obviously GNU awk has solved "1" (thanks to the Arnold et al for that!) 
> but not
> "2" yet....
>
> Does the perl solution posted earlier:
>
>     perl -p0e 's%<t1>.+?</t1>%%gs'
>
> address newlines in between the begin/end pattern and, if so, how (e.g. 
> what
> does it use for the equivalent of RS, and does it allow for the 
> "RS-equivalent"
> to be an RE)?
>
> Ed.
>
Oh! I missed the last question, the best (about gawk), from the command
line 
you can only set it to a character, from the the script you can set it to
a 
"string" not regex! I'm not sure, if you may be able to do that by 
installing some additional packages. Perl experts, fill in.

Rajan