• FauxPseudo @lemmy.world
    link
    fedilink
    arrow-up
    314
    arrow-down
    4
    ·
    1 year ago

    Ok. This covers every ipv6 and ipv4 address.

    “^\s*((([0-9A-Fa-f]1,4}:){7}([0-9A-Fa-f]{1,4}:))|(([0-9A-Fa-f]{1,4:)6}(:[0-9A-Fa-f]{1,4}((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)|:))|(([0-9A-Fa-f]1,4}:){5}(((:[0-9A-Fa-f]{1,4}){1,2}):((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)|:))|(([0-9A-Fa-f]1,4}:){4}(((:[0-9A-Fa-f]{1,4}){1,3})((:[0-9A-Fa-f]{1,4)?:((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})):))|(([0-9A-Fa-f]{1,4:)3}(((:[0-9A-Fa-f]{1,4}){1,4})((:[0-9A-Fa-f]{1,4)0,2}:((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))|:))|(([0-9A-Fa-f]1,4}:){2}(((:[0-9A-Fa-f]{1,4}){1,5})((:[0-9A-Fa-f]{1,4)0,3}:((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))|:))|(([0-9A-Fa-f]1,4}:){1}(((:[0-9A-Fa-f]{1,4}){1,6})((:[0-9A-Fa-f]{1,4)0,4}:((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))|:))|(:(((:[0-9A-Fa-f]1,4}){1,7})((:[0-9A-Fa-f]{1,4)0,5}:((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))|:)))(%.+)?\s*$”

    • Danny M@lemmy.escapebigtech.info
      link
      fedilink
      arrow-up
      84
      arrow-down
      2
      ·
      edit-2
      1 year ago

      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).

        • azertyfun@sh.itjust.works
          link
          fedilink
          arrow-up
          6
          ·
          1 year ago

          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.

        • Danny M@lemmy.escapebigtech.info
          link
          fedilink
          arrow-up
          3
          ·
          edit-2
          1 year ago

          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)
          
          • p1mrx@sh.itjust.works
            link
            fedilink
            arrow-up
            2
            ·
            1 year ago

            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.

          • Emma_Gold_Man@lemmy.dbzer0.com
            link
            fedilink
            arrow-up
            2
            ·
            1 year ago

            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.

      • Centillionaire@kbin.social
        link
        fedilink
        arrow-up
        76
        arrow-down
        1
        ·
        1 year ago

        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.

        • stoy@lemmy.zip
          link
          fedilink
          arrow-up
          13
          ·
          1 year ago

          So add two more octets:

          Moat companies will still just use something like 10.0.13.37.0.1

            • dan@upvote.au
              link
              fedilink
              arrow-up
              1
              ·
              1 year ago

              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.

              • dohpaz42@lemmy.world
                link
                fedilink
                English
                arrow-up
                18
                arrow-down
                1
                ·
                1 year ago

                My dude, you used the 10.xx private IP as an example. Why wouldn’t they assume you were referring to internal networks?

                • stoy@lemmy.zip
                  link
                  fedilink
                  arrow-up
                  1
                  arrow-down
                  3
                  ·
                  1 year ago

                  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.

        • alienzx@feddit.nl
          link
          fedilink
          English
          arrow-up
          5
          arrow-down
          1
          ·
          edit-2
          1 year ago

          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

      • Nalivai@discuss.tchncs.de
        link
        fedilink
        arrow-up
        17
        ·
        1 year ago

        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

      • Patches@sh.itjust.works
        link
        fedilink
        arrow-up
        19
        ·
        edit-2
        1 year ago

        Made that joke in an interview once.

        They didn’t think it was funny. They truly thought Regex was the solution to, but never the cause of, all problems.

        They wanted to make a Regex to verify every single address in the world. Dodged a bullet

        • rob64@startrek.website
          link
          fedilink
          arrow-up
          9
          ·
          edit-2
          1 year ago

          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.

          • Patches@sh.itjust.works
            link
            fedilink
            arrow-up
            16
            ·
            edit-2
            1 year ago

            You wouldn’t. It’s not possible. Which is what I told them.

            And why would you want to? Legally if you change the given address, and it fails to get delivered - that is on you. Not them.

            Some countries have addresses that are literally ‘Last house on the left by the Big Tree. Bumban(Neighborhood). NN (Country)’. Any US Centric validation would fail this but I assure you - mail gets delivered just fine.

            • azertyfun@sh.itjust.works
              link
              fedilink
              arrow-up
              4
              ·
              1 year ago

              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.

              • Natanael@slrpnk.net
                link
                fedilink
                arrow-up
                4
                ·
                1 year ago

                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.

    • Mr_Dr_Oink@lemmy.world
      link
      fedilink
      arrow-up
      6
      ·
      1 year ago

      Remember, when we abbreviate an ipv6 address all leading zeros are reduced to a single 0.

      E.g

      0003 would just become 03

      When there are geoups of 4 zeros these can be represented as a single 0 or as a double colon ::

      But we can only use the :: once so when summarizing an address containing multiple groups of 4 0s one after the other they can all be abbreviated to a single ::

      Eg

      fe80:0000:0000:0000:0210:5aff:feaa:20a2 would become fe80::210:5aff:feaa:20a2

      Therefore it is perfectly valid to abbreviate an address of 0000:0000:0000:0000:0000:0000:0000:0000 /0 to just ::/0

      • Static_Rocket@lemmy.world
        link
        fedilink
        English
        arrow-up
        2
        ·
        1 year ago

        Eh, I’ve seen some software internally prefer 0::0 instead of just ::0 or :: . Notation wise though you are correct, it is unnecessary.

      • Mr_Dr_Oink@lemmy.world
        link
        fedilink
        arrow-up
        2
        ·
        1 year ago

        Its CIDR notation. So /0 means the subnet mask has no on bits and would read as 0.0.0 0 if you had a /1 that turns 1 bit on in the subnet mask, so it would be 128.0.0.0.

        If i had a /24 which is the subnet mask used for most small networks like your home router. There would be 255 minus 2 addresses available for clients (phones, pcs etc) so the subnet mask would have 24 on bits and read 255.255.255.0, which you may be familiar with.

        (Assuming you dont know much, not to insult you, you might know plenty), but when writing any kind of instructions or guides, i was always told to assume the reader knows absolutely nothing and miss nothing out.

  • dan@upvote.au
    link
    fedilink
    arrow-up
    27
    ·
    edit-2
    1 year ago

    This reminds me of something I saw online maybe 20 years ago now. Someone created a torrent with a name like “every IP address ever (hacking tool)” and uploaded it to Suprnova, which ended up having thousands of people seeding it. It was just a text file with every IPv4 from 0.0.0.0 to 255.255.255.255 😂

      • spuncertv@iusearchlinux.fyi
        link
        fedilink
        arrow-up
        10
        ·
        edit-2
        1 year ago

        That file would be ungodly large. There are 2^128 possible addresses, each weighing in at 128 bits, 16 bytes. 16 bytes times 340 trillion trillion trillion. That puts us around 5.44 trillion Zettabytes. The estimates I’ve seen for worldwide data storage sit aroun 60-70 zettabytes.

      • dan@upvote.au
        link
        fedilink
        arrow-up
        5
        ·
        edit-2
        1 year ago

        IPv6 version is just a Python script that generates random 128-bit integers. Eventually you’ll hit a valid IPv6 address!