IRC logs for #openrisc Thursday, 2012-04-12

juliusb_jonibo_: DSX implementation done I believe :)07:46
jonibo_juliusb_: I just saw the patch!  best looking thing I've seen go across the wire yet!07:58
jonibo_juliusb_: that new key should work for you now, too... give it a try and let me know if it doesn't07:58
juliusb_cool, will have a crack08:05
juliusb_hmm, key not working08:08
juliusb_in my ~/.ssh/config I have :08:08
juliusb_Host openrisc.net Hostname openrisc.net User julius IdentityFile ~/.ssh/id_rsa08:08
juliusb_argh, with \n after much of that08:08
jonibo_hehe08:08
jonibo_not sure how to interpret all your comment there... not working?08:09
juliusb_yeah, i still get fatal: The remote end hung up unexpectedly08:09
jonibo_ok08:09
juliusb_that's maybe not an auth issue?08:10
jonibo_try again... i'm not seeing any login attempt from you08:11
jonibo_i see the login attempt...08:12
juliusb_ok i added a new origin with ssh://julius@openrisc.net.....08:12
juliusb_and just pushed to that08:12
juliusb_and it's there08:12
jonibo_great08:12
jonibo_so the key works08:12
jonibo_perfect08:12
juliusb_yeah, awesome, thanks for that08:12
jonibo_what was your old origin?08:12
juliusb_just an anon checkout I presume08:13
juliusb_i cloned it to apply the patches today onto this machine08:13
jonibo_yeah, ok08:13
jonibo_makes sense08:13
jonibo_juliusb_: any chance you could try pushing to code@openrisc.net:openocd  ?08:19
jonibo_just for kicks08:19
juliusb_the whole repo?08:20
jonibo_yeah08:20
juliusb_ok08:20
juliusb_how do you recommend I d oit?08:21
juliusb_should I add a new remote?08:21
jonibo_git push code@openrisc.net:openocd master08:22
juliusb_ok08:22
jonibo_nah, you don't need to add a new remote right now... just the simple push is fine for now08:22
juliusb_i Added --all08:22
jonibo_ok08:22
juliusb_to get all the branches08:22
jonibo_i'm just interested to know that it works for you08:23
juliusb_yep i'm writing08:23
jonibo_ok, good08:23
jonibo_i'll start working on the cleanup that I talked about... that new URL will be the new repo08:23
jonibo_i think it's cleaner08:23
jonibo_and you can give write access to others to that same repo, if you so desire08:24
jonibo_i'll leave it up to you08:24
jonibo_but I'll get back to you when I'm done with the cleanup08:24
juliusb_ok nice08:26
juliusb_oh my typo of "adv_debug_syys" showed up :(08:26
juliusb_fixed08:29
stekernthat's the finnish version? ;)08:35
juliusb_wouldn't that be aiuuudv_deeeeiubuuug_syys?08:37
jonibo_after a few drinks, yes08:38
juliusb_:)08:39
stekernhehe, probably, but aiuuudv_deeeeiubuuug doesn't mean anything opposed to syys, which means "autumn" :)08:42
juliusb_haha, I guess autumn does debug things a bit by making all of the flowers and trees lose their leaves08:43
juliusb_nowhere for the bugs to live08:44
stekernpoetic ;)08:44
stekernand for the drunk part, that's the default state for finnish people08:44
juliusb_not for the bugs, they lose their homes! But maybe that's when they metamorphose and migrate into code databases08:45
juliusb_I presume that drinking is due to the proximity of Russia08:46
stekernyes, that, and bad weather08:47
stekernjonibo_: I spent last evening looking through the llvm-openrisc backend, your tablegen instrinfo is far ahead of the one in there08:54
jonibo_really?08:54
jonibo_ok, that's disappointing08:54
jonibo_i guess 24 hours didn't quite cut it then! :)08:55
jonibo_pretty sure I spent more than 24 hours on that backend already...08:55
stekernyeah, he basically just have l.nop, l.add{i}, l.or{i} and return (l.jr r9)08:56
stekernand no delayslots (although I hacked a simple one just filling with l.nops yesterda)08:57
stekernI'm starting to get my head (slightly) around the tablegen stuff at least08:57
stekernwas pretty confusing at first glance, but it seems really nice when you get to know it08:58
jonibo_yeah, that's the thing... it's brutal at first glance, but pretty well thought-out once you start poking at it08:59
stekernI threw the few lines needed in clang too, so I was able to compile a simple int foo(void){return 0;} function ;)09:00
jonibo_nice!09:00
stekernI like that you have a well defined language that you are translating against, how does that work in gcc?09:03
jonibo_gcc has that too09:10
jonibo_it's pretty much GIMPLE09:10
jonibo_in fact, GCC keeps GIMPLE in the object files now in order to allow LTO (link time optimization)09:11
jonibo_LLVM does this with it's IR directly09:11
stekernI see09:12
jeremybennettstekern: jonibo: Good work on LLVM. Where can I pull your latest work from?10:43
stekernjeremybennett: I was trying out the work by anton yesterday10:44
jeremybennettFrom the comments above, I gather jonibo's work is rather ahead of that.10:45
stekernah, yes, at least in the tablegen instrinfo department10:46
jeremybennettI might try to merge the best of both together to keep me busy while I'm at this conference.10:46
stekernthat one is still available at: https://github.com/skristiansson/llvm-or1k10:46
jeremybennettThanks - I'll pull it over.10:46
jeremybennettI see Anton K's stuff still uses the or32 notation, so it would be good to move him over to or1k.10:47
stekernare you at the llvm conference now?10:47
jeremybennettIt starts at noon, and I'll be there.10:47
stekernyes, I noted that too, it's a minor issue10:47
stekerncool, be sure to report back here ;)10:48
jeremybennettI'll try to make sure I spend time with Anton K. Will report back.10:48
unicosdonated to the ASIC project last night, cheers!21:28

Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!