IRC logs for #openrisc Saturday, 2015-01-17

--- Log opened Sat Jan 17 00:00:59 2015
olofkpoke53282: I just used jor1k to solve a problem :)06:12
olofkOh, just remembered. Should we apply for GSoC this year?06:15
stekernolofk: when is that?06:31
stekernwe should probably start preparing a bit earlier than last year06:31
olofkstekern: Yeah, I can't remember the dates any more. I read something a few months ago, but kept forgetting to look it up06:41
olofkok, we can start applying as an org February 906:44
olofkDid we colllect any proposals last year?06:45
olofkproject proposals, I mean06:45
stekernyes, there was some wiki page blueCmd put up iirc06:52
olofk"Some wiki page". This seems to be the case of OpenRISC documentation in general08:15
poke53282olofk: At the moment, around 60 people are visiting the website each day. At least the people who allow google analytics.11:34
poke53282But there are sometimes pretty significant peaks.11:34
poke53282So, unfortunately not so many people are using this site to solve a problem.11:35
poke53282But I guess the real number of visitors is higher.11:38
poke53282Because most of the people who are interested in such a project would also have addons to ban trackers.11:39
stekernthat comment made me think about a pretty interesting swedish documentary about stolen identities on internet12:31
stekernthe conclusion they made was that the more promiscous you were, the less chance the "identity stealer" would have to seem like they are you12:31
stekerniow, your own strong precense on internet overshadows them impersonating you12:34
olofkThe problem with that is you need to be visible in many places on the internet13:20
olofkSo for me, someone could probably successfully steal my identity as long as they stay away from the opencores bugtracker or #openrisc :)13:20
stekernnah, you're a little twitter bitch, you'll stay safe13:21
stekernit's worse for me13:21
olofkNo problem for you either. We would quickly see the drop in progress on everything OpenRISC-related and ring the alarm :)13:22
stekernthis was the documentary btw: https://www.youtube.com/watch?v=6JFIfvZV2VM13:28
olofkjeremybennett: I'm trying to kill off old bugs in OpenCores bugzilla, and got some questions for you. Are you around?13:29
olofkstekern: Cool. I got to watch that13:30
olofkFor heaven's sake. How long time can it take to compile a compiler? It's been running for more than three hours now14:24
ysangkokpoke53282: i'm having some problems with jor1k-toolchain... the version of libffi referenced is older than 3.2.1, so i assume it doesn't support openrisc yet14:50
ysangkokpoke53282: so i tried updating the version number in the make file...14:50
ysangkokpoke53282: but now the include file is installed in /usr/lib/libffi-3.2.1 which of course doesn't work14:51
ysangkokpoke53282: well, i assume it can't work. but the actual error i get is /opt/cross/or1k-linux-musl/lib/gcc/or1k-linux-musl/4.9.0/../../../../or1k-linux-musl/bin/ld: cannot find -lffi14:53
ysangkok(when building glib)14:53
ysangkokor, rather, linking14:53
ysangkokah, i see libffi uses pkg-config14:55
ysangkoklooks like somehow  -L${toolexeclibdir} from the pkg-config isn't getting written.. maybe it's the pkg-config-path... i'll research14:59
stekerndown to 83 gcc failures now16:10
olofkstekern: Sounds like it's soon time for me to send you a Steam gift code ;)16:45
olofkCool. Seems like the newlib guys are positive to doing a 2.2.1 release with the last or1k patches17:07
olofkBut they failed to build or1k-gcc. Haven't had any problems myself. Anyone else got any clues?17:09
olofkhttps://sourceware.org/ml/newlib/2015/msg00042.html17:09
poke53282ysangkok: Yes my libffi script might be wrong. This is the script for my version of libffi before I send the patches.17:29
poke53282even an update of the version string might not work as you said.17:30
poke53282do you use my scripts? In the Makefile I set several environment variables for pkg-config17:30
poke53282stekern: If you send me your Steam ID maybe I have something for you as well. :)17:31
poke53282stekern: Do you have a repository with updates?17:34
stekernthe only "real" gcc bug so far is in or1k-gcc17:49
stekernother have just beeen musl related17:49
stekernlike profiling tests needed to be disabled17:50
poke53282I guess one is related to a missing stack usage function.18:12
olofkNoooo!!! make install failed :(18:17
olofkmkdir -p -- /usr/local/libexec/gcc/i686-pc-linux-gnu/4.9.1/install-tools18:18
olofkmkdir: cannot create directory ‘/usr/local/libexec’: Permission denied18:18
olofkYou got to be kidding me. I found a bug in my makefile so I have spent the day compiling a native version to be installed in /usr/local18:19
olofkA ':=' instead of '+=' reset the configure arguments :(18:20
poke53282:)18:36
poke53282olofk: do you a raspi to compile?18:37
olofkpoke53282: Almost. It's an Atom N27022:17
dalias:)22:41
daliasatoms are not that slow22:41
poke53282That explains it.22:49
poke53282Well, I had one by my own. And I installed gentoo on it.22:49
poke53282I regret it22:50
dalias:)22:51
poke53282olofk: Use the options "--enable-languagues=c"23:24
poke53282add only a comma separated list of programming languages, which you need.23:24
poke53282=c,c++23:24
poke53282maybe23:24
olofkpoke53282: Yeah, I'm using c,c++ but I'm considering using only c next time. Not sure I ever used c++ for this tc23:34
olofkAnd I'm running gentoo on this atom. Perfectly happy with that. Just got firefox as a prebuilt package23:35
olofkGot ubuntu installed on it as well, and that is insanely slow compared to the gentoo system23:36
olofkI see it as Amdahls law. Upgrading is slower, but using the system is faster, and I spend much more time using the system than upgrading it23:38
--- Log closed Sun Jan 18 00:00:00 2015

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