note ignatz So true... within each response one can see a little window into the responders soul: <dl> <dt><b>Angry monk</b>:</dt> <dd>[id://9953] [id://17316] [id://120486] [id://20443] [id://172086] [id://78752] [http://www.tuxedo.org/~esr/faqs/smart-questions.html|How to ask questions the smart way]</dd> <dt><b>All forrest no trees monk</b>:</dt> <dd>[use strict or die];</dd> <dt><b>Curt monk</b>:</dt> <dd>see [pacopod]</dd> <dt><b>I already asked that monk</b>:</dt> <dd>I already asked that [paco|here].</dd> <dt><b>Too much time on his hands monk</b>:</dt> <dd>[Here] [are] [several] wonderful nodes on the subject by our beloved brother [tilly] that I have found useful.</dd> <dt><b>No time on his hands monk</b>:</dt> <dd>Don't reinvent the wheel, use [cpan://God::Paco] from the CPAN.</dd> <dt><b>Too helpful monk</b>:</dt> <dd>As you see in the code below, I've rewritten your regex and incorporated a rather interesting example of the Swartzian Transform. Also, I've made it object-oriented to improve readability. </dd> <dt><b>Module monk</b>:</dt> <dd>I've written a [pacopod|module] for this that I haven't published to the CPAN yet.</dd> <dt><b>Column monk</b>:</dt> <dd>I've already written a column about that.</dd> <dt><b>Benchmark monk</b>:</dt> <dd>Your solution is very slow compared to Perl's native paco function: <code>Benchmark: timing 10000 iterations of paco, wako... paco: 4 secs ( 3.97 usr 0.01 sys = 3.98 cpu) wako: 6 secs ( 4.97 usr 0.00 sys = 4.97 cpu)</code></dd> <dt><b>Einstein monk</b>:</dt> <dd><code> my $contextual = '(?:c(?:bex|e(?:bax|vpxf?)|hff(?:vrf|l)|vff(?:\\ gnxr|\\-gnxr|gnxr|r(?:ef|[feq])|vat|l)?)|dhvzf?|ebbg(?:r(?:ef|[eq])|vat|f)?|f(?:bq(?:q(?:rq|vat)|f)?|chax|perj(?:rq|vat|f)?|u(?:nt(?:t(?:r(?:ef|[qe])|vat)|f)?|vg(?:g(?:r(?:ef|[qe])|vat|l)|r(?:ef|[fqel])|vat|[fr])?))|g(?:heqf?|jngf?|vgf?)|jnax(?:r(?:ef|[eq])|vat|f)?|n(?:ef(?:r(?:\\ ubyr|\\-ubyr|ubyr|[fq])|vat|r)|ff(?:\\ ubyrf?|\\-ubyrf?|rq|ubyrf?|vat))|o(?:ba(?:r(?:ef|[fe])|vat|r)|h(?:ttre|yy(?:\\ fuvg(?:g(?:r(?:ef|[qe])|vat)|f)?|\\-fuvg(?:g(?:r(?:ef|[qe])|vat)|f)?|fuvg(?:g(?:r(?:ef|[qe])|vat)|f)?))|n(?:paco|yy(?:r(?:ef|[qe])|vat|f)?)|yb(?:bql|j(?:\\ wbof?|\\-wbof?|wbof?)))|p(?:bpx(?:\\ fhpx(?:ref?|vat)|\\-fhpx(?:ref?|vat)|fhpx(?:ref?|vat)|f)?|enc(?:c(?:r(?:ef|[eq])|vat|l)|f)?|h(?:agf?|z(?:vat|zvat|f)))|q(?:batf?|vpx(?:\\ urnq|\\-urnq|rq|urnq|vat|yrff|f)?)|s(?:hpx(?:rq|vat|f)?|neg(?:r[eq]|vat|[fl])?|rygpu(?:r(?:ef|[efq])|vat)?)|u(?:hzc(?:r(?:ef|[eq])|vat|f)?|n(?:eq[\\-\\ ]?ba|ys(?:\\ n[fe]|\\-n[fe]|n[fe])frq))|z(?:bgure(?:\\ shpx(?:ref?|vat)|\\-shpx(?:ref?|vat)|shpx(?:ref?|vat))|hgu(?:n(?:\\ shpx(?:ref?|vat|[nnn])|\\-shpx(?:ref?|vat|[nnn])|shpx(?:ref?|vat|[nnn]))|re(?:\\ shpx(?:ref?|vat)|\\-shpx(?:ref?|vat)|shpx(?:ref?|vat)))|reqr?))';</code></dd> </dl> <p> (UPDATE: Fixed some thread display issues. Funny how an HTML definition list looks different in a thread than as a stand alone node.)<P> (UPDATE 2: Doing an XHTML version of a DL seems to fix the formatting on the different views (Thanks to [Aristotle] for the idea). Let me know if there are still problems. Also, how about golf monk? ) <pre style="font-family: terminal, courier, courier new, monospaced;"> ()-() \"/ ` </pre> 206653 206656