Please don’t. Use regex to find something that looks like an IP then build a real parser. This is madness, its’s extremely hard to read and a mistake is almost impossible to spot. Not to mention that it’s slow.
Just parse [0-9]{1,3}.[0-9]{1,3}.[0-9]{1,3}.[0-9]{1,3} using regex (for v4) and then have some code check that all the octets are valid (and store the IP as a u32).
Fuck that, if for whatever reason I’m writing an IP validator by hand I’m disallowing leading zeros. Parsers are very inconsistent, some will parse 010 as 10, others as 0o10 == 8 (you can try that right now with a POSIX ping). Talk about a footgun.
Because 1.2.3.4 and 1.02.003.04 both map to the same number.
But 10.20.30.40 and 010.020.030.040 map to different numbers. It’s often best to reject IPv4 addresses with leading zeroes to avoid the decimal vs. octal ambiguity.
True enough for database or dictionary storage, but a lot of times things get implemented in arrays where you still wind up with two copies of the same uint32.
That would allow for like, 2 trillion devices? Feels like a bandaid, my dude. Next you’re gonna suggest a giant ice cube in the ocean once a year to stop global warming.
You can use a ULA if you want to. That’s essentially the IPv6 equivalent of a private IP.
Why though? Having the same IP for both internal and external solves a bunch of issues. For example, you don’t need to use split horizon DNS any more (which is where a host name has a different IP on your internal network vs on the internet). You just need to ensure your firewalls are set up properly, which you should do anyways.
I thought it was pretty clear with me adding 13.37 that I was making a joke, the earlier post spoke about how just adding one octet would still be too few addresses, so I joked about adding one more octet.
Hurricanes cannot cross the equator. The equator is an imaginary line, and hence has zero mass. We can end every hurricane using zero point zero energy (0.0).
You could follow this logic and add 2 alphanumeric digits before 4 numeric octets. E.g. xf.192.168.1.1
This would at least keep it looking like an IP and not a Mac address. Another advantage would be graceful ipv4 handling with a reserved range starting with “ip” like ip.10.10.10.1
Oh yeah, great, let’s change the fundamental protocol on which all the networks in the world are based. Now two third of the devices in the world crashed because you tried to ping 192.168.0.0.1
Holy hell yeah you did. How would you go about doing that in a single expression? A bunch of back references to figure out the country? What if that’s not included? Oy.
The only valid regex is (.+). Maybe add a separate country field (especially because some Americans wholeheartedly believe that the entire world should understand that “foobar, TX” means “foobar, Texas, United States”) (don’t get me started on states whose abbreviations are also ISO country codes).
Unfortunately I guess business people only care about getting fewer support calls for missing shipping details, not correctness or a couple of calls from customers who live in the boonies. Then the proper answer is a form with a bunch of fields… which Americans will inevitably fuck up by making the “State” field mandatory despite most countries not having an equivalent.
What I’d really do is use one of those services that automatically fill on the address using google maps or whatever. Not perfect, probably not free, but a whole lot less work for presumably way fewer PEBCAKs from customers.
If you’re using one of those services then PLEASE allow manual entry / override because I’ve had forms like that which I were blocked from filing in because it didn’t acknowledge that my address existed.
Ok. This covers every ipv6 and ipv4 address.
“^\s*((([0-9A-Fa-f]1,4}:){7}([0-9A-Fa-f]{1,4}:)6}(:[0-9A-Fa-f]{1,4})|:))|(([0-9A-Fa-f]1,4}:){5}(((:[0-9A-Fa-f]{1,4}){1,2}))|:))|(([0-9A-Fa-f]1,4}:){4}(((:[0-9A-Fa-f]{1,4}){1,3}))?:((25[0-5]|2[0-4][0-9]|1[0-9][0-9]|[1-9]?[0-9])(.(25[0-5]|2[0-4][0-9]|1[0-9][0-9]|[1-9]?[0-9]))3})):)3}(((:[0-9A-Fa-f]{1,4}){1,4}))0,2}:((25[0-5]))|:))|(([0-9A-Fa-f]1,4}:){2}(((:[0-9A-Fa-f]{1,4}){1,5}))0,3}:((25[0-5]))|:))|(([0-9A-Fa-f]1,4}:){1}(((:[0-9A-Fa-f]{1,4}){1,6}))0,4}:((25[0-5]))|:))|(:(((:[0-9A-Fa-f]1,4}){1,7}))0,5}:((25[0-5]))|:)))(%.+)?\s*$”
deleted by creator
Please don’t. Use regex to find something that looks like an IP then build a real parser. This is madness, its’s extremely hard to read and a mistake is almost impossible to spot. Not to mention that it’s slow.
Just parse [0-9]{1,3}.[0-9]{1,3}.[0-9]{1,3}.[0-9]{1,3} using regex (for v4) and then have some code check that all the octets are valid (and store the IP as a u32).
And dupe check. 0.0.0.0 and 000.000.000.000 may both be valid, but they resolve the same
Fuck that, if for whatever reason I’m writing an IP validator by hand I’m disallowing leading zeros. Parsers are very inconsistent, some will parse 010 as 10, others as 0o10 == 8 (you can try that right now with a POSIX
ping). Talk about a footgun.…and that’s me in the fetal position, thanks.
Definitely, tho if you store it as a u32 that is fixed magically. Because 1.2.3.4 and 1.02.003.04 both map to the same number.
What I mean by storing it as a u32 is to convert it to a number, similar to how the IP gets sent over the wire, so for v4:
octet[3] | octet[2] << 8 | octet[1] << 16 | octet[0] << 24
or in more human terms:
(fourth octet) + (third octet * 256) + (second octet * 256^2) + (first octet * 256^3)But 10.20.30.40 and 010.020.030.040 map to different numbers. It’s often best to reject IPv4 addresses with leading zeroes to avoid the decimal vs. octal ambiguity.
I don’t know why anyone would write their IPs in octal, but fair point
It’s not about how people write them, it’s how parsers parse them. IPv4 has been around since 1982, and most parsers interpret leading zeros as octal.
True enough for database or dictionary storage, but a lot of times things get implemented in arrays where you still wind up with two copies of the same uint32.
deleted by creator
IPv6 was a mistake. We should have just added an addition octet
That would allow for like, 2 trillion devices? Feels like a bandaid, my dude. Next you’re gonna suggest a giant ice cube in the ocean once a year to stop global warming.
So add two more octets:
Moat companies will still just use something like 10.0.13.37.0.1
IPv6 is not made with internal networks in mind lol
You can use a ULA if you want to. That’s essentially the IPv6 equivalent of a private IP.
Why though? Having the same IP for both internal and external solves a bunch of issues. For example, you don’t need to use split horizon DNS any more (which is where a host name has a different IP on your internal network vs on the internet). You just need to ensure your firewalls are set up properly, which you should do anyways.
Never claimed it was, please quote me where I said as much
My dude, you used the 10.xx private IP as an example. Why wouldn’t they assume you were referring to internal networks?
I thought it was pretty clear with me adding 13.37 that I was making a joke, the earlier post spoke about how just adding one octet would still be too few addresses, so I joked about adding one more octet.
I’m only pointing out why the other poster would make the assumption you were referring to an internal network. Do with it what you will.
And nuke the hurricanes
Hurricanes cannot cross the equator. The equator is an imaginary line, and hence has zero mass. We can end every hurricane using zero point zero energy (0.0).
o.O
ONCE AND FOR ALL
AND MY AXE!
You could follow this logic and add 2 alphanumeric digits before 4 numeric octets. E.g. xf.192.168.1.1
This would at least keep it looking like an IP and not a Mac address. Another advantage would be graceful ipv4 handling with a reserved range starting with “ip” like ip.10.10.10.1
Oh yeah, great, let’s change the fundamental protocol on which all the networks in the world are based. Now two third of the devices in the world crashed because you tried to ping 192.168.0.0.1
that WOULD be quite funny for the first second or 2…
Could have sped up adoption significantly.
They played us for absolute fools!
heared of ipv5?
Plus the MAC address
https://imgs.xkcd.com/comics/perl_problems.png
deleted by creator
Holy hell yeah you did. How would you go about doing that in a single expression? A bunch of back references to figure out the country? What if that’s not included? Oy.
deleted by creator
The only valid regex is
(.+). Maybe add a separate country field (especially because some Americans wholeheartedly believe that the entire world should understand that “foobar, TX” means “foobar, Texas, United States”) (don’t get me started on states whose abbreviations are also ISO country codes).Unfortunately I guess business people only care about getting fewer support calls for missing shipping details, not correctness or a couple of calls from customers who live in the boonies. Then the proper answer is a form with a bunch of fields… which Americans will inevitably fuck up by making the “State” field mandatory despite most countries not having an equivalent.
What I’d really do is use one of those services that automatically fill on the address using google maps or whatever. Not perfect, probably not free, but a whole lot less work for presumably way fewer PEBCAKs from customers.
If you’re using one of those services then PLEASE allow manual entry / override because I’ve had forms like that which I were blocked from filing in because it didn’t acknowledge that my address existed.
/.+(road|street).+/ resigns
It’s always a treat to debug a regex of that size.
I knew there would be someone with the regex.
You’re more of a perl programmer than network engineer :P
.*
Technically, this one also matches everything:
*exits the room*