strange behaviour: continue with \G and /gc and look forward

leszekdubiel has asked for the wisdom of the Perl Monks concerning the following question:

Dear Perl Monks! Please explain why the program doesn't print "Z: passed" but displays "E: failed"?

#!/usr/bin/perl


my $a = " -1";  

# remove trailing spaces 
warn "A: pos is ", pos $a;   
$a =~ /\G */gc;

warn "B: pos is ", pos $a;    
$a =~ /\G(?=-)/gc or die "C: failed";     

warn "D: pos is ", pos $a;
$a =~ /\G(?=-)/gc or die "E: failed";      

print "Z: passed";
[download]

output is:

 
A: pos is  at ./l2 line 7.
B: pos is 1 at ./l2 line 10.
D: pos is 1 at ./l2 line 13.
E: failed at ./l2 line 14.
[download]

Strange is that "pos $a" is 1, and look ahead for minus sign fails. B: situation is exactly the same as D:, but C: doesn't fail, but E: does...

Comment on strange behaviour: continue with \G and /gc and look forward Select or Download Code

Replies are listed 'Best First'.
Re: strange behaviour: continue with \G and /gc and look forward by tybalt89 (Monsignor) on Dec 13, 2019 at 21:12 UTC
`The higher-level loops preserve an additional state between ite +rations: whether the last match was zero-length. To break the loop, the following match after a zero-length match is prohibited to have + a length of zero. This prohibition interacts with backtracking ( +see "Backtracking"), and so the second best match is chosen if the +best match is of zero length.` [download]	[reply] [d/l]
Re^2: strange behaviour: continue with \G and /gc and look forward by leszekdubiel (Scribe) on Dec 13, 2019 at 21:35 UTC
Ok thank you very much. Community around perl is great. Actually I've been programming perl for a long time and this bite me for the first time. Here is the link: https://perldoc.perl.org/perlre.html#Backtracking search for phrase " higher-level loops "	[reply]
Re: strange behaviour: continue with \G and /gc and look forward by tybalt89 (Monsignor) on Dec 13, 2019 at 21:04 UTC
regex will not match twice at the same position, it will advance by one on the second attempt. It's somewhere in the docs, but I can't find it just now.	[reply]
Re^2: strange behaviour: continue with \G and /gc and look forward by ikegami (Patriarch) on Dec 14, 2019 at 17:33 UTC
It can match at the same position twice. `say "[$-[0], $+[0])" while "aaaaa" =~ /a*?/g;` [download] `[0, 0) [0, 1) [1, 1) [1, 2) [2, 2) [2, 3) [3, 3) [3, 4) [4, 4) [4, 5) [5, 5)` [download]	[reply] [d/l] [select]
Re^3: strange behaviour: continue with \G and /gc and look forward by tybalt89 (Monsignor) on Dec 14, 2019 at 17:52 UTC
I should have said "regex will not ZERO LENGTH match twice at the same position".	[reply]
Re: strange behaviour: continue with \G and /gc and look forward by ikegami (Patriarch) on Dec 14, 2019 at 17:43 UTC
The following can be used to reset the flag: `pos($a) = pos($a);` [download]	[reply] [d/l]
Re: strange behaviour: continue with \G and /gc and look forward by rsFalse (Chaplain) on Dec 16, 2019 at 10:21 UTC
Maybe this prose also relates here: Repeated Patterns Matching a Zero length Substring (updated a link, thanks AnomalousMonk and soonix for pointing)	[reply]
Re: strange behaviour: continue with \G and /gc and look forward by rsFalse (Chaplain) on Dec 16, 2019 at 10:30 UTC
>>$a and $b are special variables which I'd recommend avoiding except for their special purposes. See perlvar: $a for more information. -- kcott (cite from Re^3: Switch/Format to not interpret metacharacters in regex?)	[reply]