Using chained comparator

[habere-et-dispertire]

Does this read better ?

-$/.Int <= 255 && $/.Int >= 0
+0 <= $/.Int <= 255

[coke]

I think so, yes.

[raiph]

0 <= $/.Int <= 255

I don't know the context, intent of the example, etc, but why not 0 <= $/ <= 255?

Is there some numeric coercion edge case that I'm not taking into account but would occur in the intended context? (eg if $/ was ｢255.6｣ then the .Int coercion would make a difference because .Int rounds down to 255, so 0 <= $/.Int <= 255 would be True, but 0 <= $/ <= 255 would be False).

[habere-et-dispertire]

my regex ipv4-octet { \d ** 1..3 <?{ 0 <= $/ <= 255 }> }

It looks to me like the match could only be a number comprising of 1 to 3 digits and so an implicit cast is safe ?

FWIW if I run the full code snippet, then I get the same result.

The next example negates things and so things get split up:

$o < 0 || $o > 255

So perhaps it would be even more clearer to use :

$/ ~~ 0..255
$/ !~~ 0..255

and make them more self-similar instead ?

[raiph]

Ah. OK. Yeah. But a few things to note:

• I was assuming the issue here was some code showing an example of chained comparison. If not, then fair enough.

• Regardless of anything else, at least for parsing an ip octet (or ip address) it should probably be a token rather than a regex. (Rakoons, and thus examples, ought generally use token as the rule kind they default to using, not regex. And if I understand what's intended with this example, a token would fail faster for many invalid inputs, and thus have better performance.)

• The token (or regex) needs to be carefully anchored. Consider, for example:

my token ipv4-octet { \d ** 1..3 <?{ $/ ~~ 0..255 }> }
say '301' ~~ &ipv4-octet; # ｢30｣  <-- Uhoh

• In practice one would likely write a token (or regex) that deals with a whole ip address in one go. Perhaps any example ought follow suit. Then again, it's definitely getting very complicated for a simple example:

my token ipv4 { ^ (\d ** 1..3) ** 4 % '.' <?{ $0.all ~~ 0..255 }> $ }

say '192.168.0.0.' ~~ &ipv4

[habere-et-dispertire]

The document is about regexes, so not sure if going into token land is on topic. It feels like non-trivial changes needs liaison with the primary author. I was just making a small suggestion while reading it.

[coke]

I agree just the small change is fine. Are you happy with the PR as is? @raiph can submit a separate PR for the change they are suggesting.

[habere-et-dispertire]

Yes, thank you.

[coke]

Done, thanks for the cleanup.