Conversation

Notices

  1. [sp][b]Obyqgrkg va n fcbvyre? Vg'f zber yvxryl guna lbh guvax.[/b][/sp] #

    Sunday, 22-Apr-12 00:38:11 UTC from web
    1. @bitshift [b][i][u][s][c=blue]OOpbqrnohfr Lnl![/c][/s][/u][/i][/b]

      Sunday, 22-Apr-12 00:38:56 UTC from web
    2. @bitshift Ah, excellent, so it's only the input box that BB-code tags still break in. Time to start on that, then. :D

      Sunday, 22-Apr-12 00:39:22 UTC from web
    3. @bitshift V npghnyyl svkrq gung hc, znxvat vg fb # naq # erznvarq abezny ohg grkg jnf ebgngrq cebcreyl. #

      Sunday, 22-Apr-12 00:39:40 UTC from web
      1. @minti @bitshift At least, in the post bit. The text box is still bluh. xD

        Sunday, 22-Apr-12 00:40:18 UTC from web
      2. @minti Yeah, I saw that in the change list, and tried it in the input box first, where it didn't work - I plan to fix that next. But yeah, in posts themselves, it's all good. :)

        Sunday, 22-Apr-12 00:41:33 UTC from web
        1. @bitshift If you're going to try and detect links with a regex, don't do what StatusNet did and make it span like 200 lines. LOL

          Sunday, 22-Apr-12 00:42:34 UTC from web
          1. @minti Don't worry, I only plan to catch BBcode tags. Links would probably be better to be affected by rot13 anyway, since it's entirely possible the URL itself is spoilery.

            Sunday, 22-Apr-12 00:44:26 UTC from web
            1. @bitshift True that! :p

              Sunday, 22-Apr-12 00:45:07 UTC from web
              1. @minti Not that it's impossible to catch links with a fairly concise regex (I have one in IdentiCurse that, in my extensive use, has yet to miss any valid links or catch any invalid ones), it's just that it simply isn't necessary here. :)

                Sunday, 22-Apr-12 00:46:54 UTC from web
                1. @bitshift Good, then you can avoid behemoths like this: http://gitorious.org/statusnet/mainline/blobs/master/lib/util.php#line830

                  Sunday, 22-Apr-12 00:49:58 UTC from web
                  1. @minti That... kinda makes sense, in its own strange way. They're trying to catch absolutely everything the URI spec defines as a valid URL, whereas the more sane-length regexes (such as the one I mentioned) are content to just match everything which is actually used in practice in more than a handful of cases. :)

                    Sunday, 22-Apr-12 01:02:07 UTC from IdentiCurse
    4. @izurin :3

      Sunday, 22-Apr-12 00:41:42 UTC from web