#archiveteam-bs 2014-10-20,Mon

↑back Search

Time Nickname Message
10:03 🔗 midas https://imgur.com/gallery/wGERY
10:33 🔗 schbirid yay, imgur requires javascript for "moving images" now
10:54 🔗 midas schbirid: imgur is total crap nowadays
12:22 🔗 godane so i'm past 21k in godaneinbox
15:34 🔗 godane so i decide i will be uploading Daily Tech News Show
15:34 🔗 godane looks like Tom only upload SD version of the video files without much meta data
15:35 🔗 godane https://archive.org/details/DTNS20141017_20141017
15:35 🔗 godane thats a example of the episodes he uploads
15:35 🔗 godane which is great but there is not metadata of keywords for hosts
15:44 🔗 aaaaaaaaa May find "Emulation as a Service" interesting http://bw-fla.uni-freiburg.de/
17:29 🔗 SketchCow I deal with those guys a lot.
17:30 🔗 SketchCow In summary, the non JSMESS emulation approaches are to have essentially HTML5 vnc to a machine.
17:31 🔗 aaaaaaaaa That is sort of what I was thinking it was, but thank you for the clarification.
17:33 🔗 SketchCow The machine has some powerful tools for controlling the situation, but that's all it is.
19:30 🔗 midas so because you arent able to shoot it with a tank, im guessing you are using virtualVM for getting a memory profile?
19:32 🔗 midas eclipse has a memory analyser that you can use to find what is causing the leakage so you should be able to track down more information with those tools
19:32 🔗 midas cc yipdw
19:56 🔗 yipdw midas: it's not a memory leak
19:56 🔗 yipdw (although there are those too)
19:56 🔗 yipdw it's unreleased connections, or at least that's what it appears to be
19:59 🔗 midas oh.
19:59 🔗 midas strace it?
20:00 🔗 yipdw stracing the JVM doesn't really do much
20:00 🔗 yipdw it's not unreleased connections to the database
20:00 🔗 yipdw it's the pool's internal bookkeeping
20:00 🔗 midas oh crap yeah true
20:01 🔗 midas another reason i hate tomcat :/
20:01 🔗 yipdw that is actually a wise decision
20:01 🔗 yipdw establishing a connection for every checkin/checkout takes substantial time
20:01 🔗 yipdw the downside is that without proper instrumentation it is fucking hard to figure out where a leak is occurring
20:03 🔗 yipdw first person to say something about a sufficiently strong type system preventing this gets thrown to the Hounds of Haskell to be eaten
20:05 🔗 chronomex oh, haskell isn't the right tool for that
20:06 🔗 chronomex you need something stronger, like Coq
20:06 🔗 yipdw ha
21:39 🔗 LD100 how can I create a collection on IA?
21:40 🔗 chronomex you need to get an IA admin to create you one
21:40 🔗 chronomex underscor or SketchCow can, maybe others in this channel
21:40 🔗 chronomex for now you can upload into the 'texts' collection and give all your things a tag that's the same
21:43 🔗 LD100 hm ok then I will upload the podcast files as one job so I dont need any collection
21:43 🔗 chronomex it's best to upload one show per item
21:44 🔗 LD100 hm ok
21:45 🔗 DFJustin podcasts should go in community audio rather than texts
21:45 🔗 chronomex ah, yes
21:52 🔗 LD100 ok is there some kind of coppy template or do I have to enter the same metadata all over again for every episode?
21:56 🔗 DFJustin http://blog.archive.org/2013/02/08/presetting-metadata-with-the-new-beta-uploader/
21:56 🔗 chronomex you can use the commandline uploader too if you want
22:00 🔗 LD100 DFJustin: hm I mean copying from the last file I uploaded
22:01 🔗 LD100 chronomex: do you have a link?
22:02 🔗 chronomex 1sec
22:03 🔗 chronomex LD100: https://github.com/mangtronix/ia-uploader
22:05 🔗 LD100 thx
22:08 🔗 DFJustin https://pypi.python.org/pypi/internetarchive is the new one
22:09 🔗 chronomex ah, that's the one
22:09 🔗 chronomex i'm not good at internet, which is very slow here anyway
23:59 🔗 SketchCow Yes, the python one is the one to do.
23:59 🔗 SketchCow And if you create a bunch, we can create a collection and ou can get it moved there.
23:59 🔗 SketchCow Hi from the air.

irclogger-viewer