[01:12] Okay, so what I am doing i this: [01:12] 1) I check which one is the next free shortcode in the new tr.im [01:12] 2) I find it in my database. [01:12] 3) I submit the url from the db. [01:13] Unfortunately the new tr.im seems to hang/break when the URL does not resolve anymore [01:13] Sometimes it will still store the URL even when it hangs, sometimes it won't [01:14] When storing an URL fails 3 times in a row, I will instead make the link go to "http://trim-old.tinyarchive.org/RANDOM-UUID" [01:14] Then I can later on set up a redirect from that URL to the old tr.im URL [01:14] eeeeeeeeew [01:14] that's horrible [01:15] but I'm sure the owners will shit brix [01:15] we aren't using urlte.am ? [01:15] => tr.im.urlte.am [01:16] Because if I tell it to go to tr.im.urlte.am/aaa and then someone else was faster at creating a new tr.im url, it will go from tri.im/aab to tr.im.urlte.am/aaa [01:16] oh, I see [01:17] So later on I will have to map which tr.im/CODE maps to which uuid [01:17] so you're going to have trim-old.tinyarchive.org/ redirect [01:17] ok cool [01:17] bit ugly, and I'm not stoked about having multiple domains of failure [01:17] but ok [01:17] you're the guy doing the work [01:19] Problem is I needed the url to be up and returning http 200 on requests [01:19] Otherwise tr.im would just hang again [01:19] ah, right [01:19] you're racing to get it loaded in [01:20] Haha yeah I set up a vm on amazon us east just so I could have low latency to heroku :D [01:22] I'm going to have enough reliable space soon to set up a proper redirector [01:22] :) [01:22] how much is it again, few hundred G? [01:23] At the moment it is pretty much exactly 500gb [01:23] I can eat that right up [01:23] have 9T coming in the mail tomorrow [01:23] Haha I bet you are checking the package tracking page very 5 seconds [01:24] yet to ship ... [01:25] Guess someone will have to write a proper redirector script [01:25] I can do it [01:25] it'll be pretty simple [01:25] can probably do it with just a shellscript, even [01:26] Redirecting bit.ly.urlte.am/XX to the URL in the db is easy [01:26] yes [01:26] Recognizing that yhoo.it.urlte.am/XX is also a bitly link is more difficult [01:26] oh? [01:27] hm [01:27] Bit.ly Pro aliases [01:27] ah, yes [01:27] I guess it oughtn't be a shellscript then [01:27] :P [01:28] https://github.com/ArchiveTeam/tinyarchive/blob/master/twitter_spritzer_import.py already contains an attempt to turn a full URL into a ("shortener", "shortcode") tuple [01:29] hmmmk [01:30] I might steal that and do the redirector in python then ;) [01:30] hehe [01:30] Oh, one more bad thing about new tr.im: it only supports lowercase URLs [01:30] wtfux [01:31] Yeah, so all uppercase tr.im urls will be lost anyways [01:32] Oh, and also url shortener kl.am closed sometime during january but we have >50% of the URLs saved [01:33] cool [01:34] SketchCow: hey, could we set up a dns entry wildcard on *.urlte.am for me? [01:34] I think I can pound this out tonight [01:34] damnit he's out of the country isn't he [03:37] does anybody want to jump on a website project? [04:35] soultcer: would you mind hooking me up with a partial dump? looking for <100M to develop my redirector against [14:16] send me maill. [18:53] chronomex: For a partial dump you can just download a few files from the full release [18:53] I put it up via http at http://helo.nodes.soultcer.com/urlteam/ so you don't have to get the torrent [18:54] thanks [18:55] Damn tr.im banned my ec2 address [18:56] They always return error 404 now, on every request [18:59] Bunch of idiots [19:00] The funny thing is, I am already seeing people hitting the new tr.im links: [19:01] 98.125.XXX.XXX - - [31/Jan/2013:12:27:15 +0000] "GET /28268123-c283-4cfc-b220-8b5c26327c42 HTTP/1.0" 200 - "http://thepiratebay.se/torrent/4501809/Anna_Song" "Mozilla/5.0 (Windows NT 6.2; WOW64) AppleWebKit/537.17 (KHTML, like Gecko) Chrome/24.0.1312.56 Safari/537.17" [20:08] haha the only response from tr.im that I get is error 404 and "Halt!" in the body [20:10] Halt! [20:10] ha [20:11] In Swedish, that means "slippery" - which would be even funnier imo [20:20] Well, guess it's time to write an unshortener for the new tr.im [20:25] So uh tr.im assigns code sequentially [20:27] Except it skips all codes with the letter e and i in it it seems [20:28] hahahah wtf [20:28] charset is [abcdfghjklmnopqrstuvwxyz0123456789] ? [20:29] urrr [21:04] aha, so any of the characters aeiou135 at the end of the code means it will be skipped [21:04] My "submit URLs" to tr.im script is getting more and more unreadable with all these stupid quirks [21:04] The only piece of code that is more fucked up must be the tr.im website itself [21:06] hahahahahahahaha [21:06] must be [21:06] how much did they pay for it, $35? [21:07] "tr.im - trim the account balance" [21:12] I bet the new owner thinks "Wow, my shortener is getting really popular so fast" [21:15] hah [21:18] haha