Re: Re^2: meaning of /o in regexes

[BrowserUk]

in reply to Re^2: meaning of /o in regexes
in thread meaning of /o in regexes

Thankyou diotalevi++. That is exactly the sort of answer I was looking for and it confirms my suspicions based on some fairly dodgey benchmarking.

No matter how hard I tried to isolate the benefits of qr//'ing or /o'ing, those benefits always seemed to disappear whenever I attempted to combine one or more pre-compiled regexes with each other or with some non-compiled stuff. In fact, I sometimes detected a penalty from using pre-compiled regexes other than stand-alone, though the differences were too small to quantify with any accuracy.

---

Okay you lot, get your wings on the left, halos on the right. It's one size fits all, and "No!", you can't have a different color.
Pick up your cloud down the end and "Yes" if you get allocated a grey one they are a bit damp under foot, but someone has to get them.
Get used to the wings fast cos its an 8 hour day...unless the Govenor calls for a cyclone or hurricane, in which case 16 hour shifts are mandatory.
Just be grateful that you arrived just as the tornado season finished. Them buggers are real work.

[diotalevi]

I looked even further and the get magic (see sv.c Perl_sv_2pv and seek to the "Regexp" section) associated with stringifying a qr regex is actually pretty cheap. I'd guess any real performance loss is just from having to compile a regex more than once which unless you are doing some monster regex... isn't all that much of an issue.

__SIG__
use B;
printf "You are here %08x\n", unpack "L!", unpack "P4", pack
  "L!", B::svref_2object(sub{})->OUTSIDE;

[BrowserUk]

The task that started me on the quest for speed was indeed a monster regex, it was also being called many times in a tight loop. As the regex was essentially repetitious, I originally thought that compiling the base regex with qr// and then using that in conjuction with a repeat count and the none repeating elements also compliled with /o might leach some benefits, but the reverse was true. I settled for programically generating the regex (using x n) into a single large regex and then compiling it with qr// (which appeared to give some slight performance benefit over /o). This was possibly due to the fact that when compiled with qr//, you can use the resultant variable directly ($string =~ $compiled_re) rather than needing to embed it within an m// operator (m/$compiled_re/). Maybe its slightly quicker to execute the former than the latter? The difference seemed significant enough to make it worthwhile..

---

Okay you lot, get your wings on the left, halos on the right. It's one size fits all, and "No!", you can't have a different color.
Pick up your cloud down the end and "Yes" if you get allocated a grey one they are a bit damp under foot, but someone has to get them.
Get used to the wings fast cos its an 8 hour day...unless the Govenor calls for a cyclone or hurricane, in which case 16 hour shifts are mandatory.
Just be grateful that you arrived just as the tornado season finished. Them buggers are real work.

[diotalevi]

You are slightly wrong: <code$qr = qr/./; $qr =~ m/$qr/; $qr =~ $qr;</code> is exactly identical. Again, some opcode info to illuminate. It also yields the slightly interesting form $text = 'abcd'; $text =~ qr/./ but then... maybe that's not quite normal ;-)

i  <@> leave[t1] vKP/REFC ->(end)
1     <0> enter ->2
2     <;> nextstate(main 5 qr.pl:1) v ->3
5     <2> sassign vKS/2 ->6
3        </> qr(/./) s ->4
-        <1> ex-rv2sv sKRM*/1 ->5
4           <> gvsv s ->5
6     <;> nextstate(main 5 qr.pl:2) v ->7

>> Here you see the $qr =~ m/$qr/ form

b     </> match() vKS ->c
-        <1> ex-rv2sv sK/1 ->8
7           <> gvsv s ->8
a        <|> regcomp(other->b) sK/1 ->b
8           <1> regcreset sK/1 ->9
-              <1> ex-rv2sv sK/1 ->a
9                 <> gvsv s ->a

>> And here you see the $qr =~ $qr form

c     <;> nextstate(main 5 qr.pl:3) v ->d
h     </> match() vKS ->i
-        <1> ex-rv2sv sK/1 ->e
d           <> gvsv s ->e
g        <|> regcomp(other->h) sK/1 ->h
e           <1> regcreset sK/1 ->f
-              <1> ex-rv2sv sK/1 ->g
f                 <> gvsv s ->g

__SIG__
use B;
printf "You are here %08x\n", unpack "L!", unpack "P4", pack
  "L!", B::svref_2object(sub{})->OUTSIDE;