| bug in stregex? [message #31532] |
Fri, 19 July 2002 13:54 |
Vapuser
Messages: 63
Registered: November 1998
|
Member |
|
|
I don't you, you tell me. Is this a bug?
IDL> tt=stregex('cdefaz',"(.*)(a|b|c)z",/extract)
IDL> help,tt
TT STRING = 'cdefaz'
My understanding of regular expressions says it's a bug. The first
'(.*)' should only match up to the 'a' and the second subexpression,
since it is *unqualified*, should handle the 'az', i.e. TT should
have two parts. tt[0] should be 'cdef' and tt[1] should be 'a.'
If my regex had been '(.*)((a|b|c)z)*', the result would be
understandible; then the 'greediness' of the (.*) regular expression
should have consumed the whole string because the 'zero or more'
qualifier applied to the second would have been satisfied, i.e. zero
matches. But in my example the engine should have failed when
attempting this match and should have backtracked two characters to
produce the output I suggest above.
Perl certainly does it this way:
% perl
$s="cdefaz";
print "s=$s\n";
@tt=($s =~/(.*)(a|b|c)z/);
foreach (@tt) { print "$_\n";}
this code produces the output:
s=cdefaz
cdef
a
i.e. $tt[0] = cdef and $tt[1] = a, as expected.
But regular expression are always somewhat mysterious. Am I missing
something here?
Comments?
whd
--
William Daffer: 818-354-0161: William.Daffer@jpl.nasa.gov
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
