JAA: Thanks for the naming suggestion re: metadata, I'm going to rely on ghtorrent to capture most of that--I need to grab the wiki/issues, issues seem especially valuable But that requires a large number of github keys and is quite throttled in comparison I was mostly wanting to make sure that my automated procedures did basically grab the bare repo and not fail horribly So thanks for checking ghtorrent censors a lot of author metadata for privacy reasons in their final 'presentable' version but I also keep the raws I think email is one of those (to prevent mass spam using their database) anyone have/know something for a cli torrent client (best rtorrent) that will download each torrent into a unique directory, easily named when automatically adding a new torrent? usecase: torrent site archiving with one directory per torrent "id" as name collisions are everywhere otherwise so i'm starting to upload this now : archive.org/details/KBS_World_News_Today-ASF-20051219 i got about 42gb of that in asf format i need to upload at least everything before 2010-10 kisspunch: Right. The wiki shouldn't be a problem as it's also a git repository; you can just git clone .../repo.wiki.git (instead of repo.git) to grab it. Issues are trickier though. The censoring is kind-of ridiculous since you can anyway get the email address of anyone who ever committed anything easily (unless they protected the address by using username@users.noreply.github.com instead). But I guess it does make sense since you can't easily get a bulk email address list then but need to actually clone the repositories. schbirid: That's possible, I've done it before. IIRC, my setup was that I had a watch directive which also set a custom flag on the new torrent plus an event listener which set the base directory for new torrents which had that flag set. This allowed me to have a normal watch directory and one which uses the torrent file's name as the target directory. I also remember that it was a nightmare to set this up. I spent days trying to figure out how rtorrent's quoting and escaping mechanisms worked. i can imagine =( do you remember how you determined the ids? right now i use autodl, i guess i would have to add it there This was all manual, i.e. I'd drop a torrent file with the right name into the warch directory. No experience with autodl. I guess it shouldn't be too difficult to set something up if you want to use the info hash as the "ID". If you want the numeric ID that's used on the tracker, it probably strongly depends on the tracker and how you download the torrents exactly. yeah, i will look into it I'll check if I still have the .rtorrent.rc from back then somewhere. Somewhere deep in .../copy-of-old-machine/copy-of-previous-machine/copy-of-broken-server/... details of the OVH outage: http://status.ovh.net/?do=details&id=15162#comment18119 :D per hash is a great idea much more useful than the internal site id !ig a72qwdi5xvc9hijjs21e2gjat ^https?://www\.msfn\.org/board/index\.php\?app=core&module=system§ion=notifications&do=followers&follow_app=forums&follow_area=topic&follow_id=(120444|128807|139877)&_infScroll=1&followerPage=\d{4}$ Whoops joepie91: Interesting, thanks. Probably means that the price increase announced by Oles will definitely come though. https://blog.cloudflare.com/cloudflare-supports-privacy-pass/ https://video.vice.com/en_us/video/the-lost-art-of-canada39s-doomed-pre-internet-web/559588b70d6551143782d428 so the buffy tape was mostly recorded over it said to have bufffy s06e06 to s06e12 but hbo sopranos for about half of the tape and the cw airings of smallville If anyone here is using the Stylish addon for Firefox: the current version 3.0.1 is horribly broken, so you might want to wait with updating. "Horribly broken" = all styles are gone (though they still exist in a file in the profile directory), and it causes 100% CPU usage for me.