| --- Log opened Wed Jan 21 00:00:06 2015 | ||
| jeremybennett | olofk: Well done with your bugzilla clean up | 15:09 |
|---|---|---|
| olofk | jeremybennett: Thanks. Hope you are ok with me closing bugs even though they might still exist in the or32 toolchain | 16:40 |
| aburgess | I'm interested in contributing patches to openrisc binutils/gcc, how should the patches be put forward? mailing list, github pull request, something else? | 16:41 |
| olofk | aburgess: We prefer fax | 16:52 |
| * aburgess digs out the old fax machine ..... wait .... you're joking :) | 16:53 | |
| olofk | :) | 16:54 |
| aburgess | I can print out the diffs and post them if that helps... | 16:54 |
| olofk | For binutils I suppose they should be sent directly upstream. stekern or blueCmd are the official maintainers IIRC so they will have the final word | 16:54 |
| olofk | GCC can be posted to the ml if they might be subject for discussion or as a pull request | 16:55 |
| aburgess | that's interesting as I'm building binutils using a version off of github.com/openrisc/or1k-src.git, though I admit I've not actually looked to see how different from upstream that is | 16:56 |
| olofk | Both binutils and newlib are upstream now, so except for gdb, we shouldn't have to use or1k-src anymore | 16:57 |
| olofk | (caveat for newlib. There were some issues that were resolved yesterday, so it might be good to wait for newlib-2.2.0-2, or try building from their CVS/git head) | 16:59 |
| aburgess | ok, thanks | 16:59 |
| aburgess | w.r.t. gdb, the expected procedure then is to build all but gdb from upstream, and gdb from the repo I mentioned earlier? | 17:00 |
| olofk | yes | 17:02 |
| aburgess | thank you | 17:02 |
| olofk | http://wallento.github.io/newlib/ | 17:02 |
| olofk | Quick'n'dirty instructions. You can change the or1k-src to binutils in the first step | 17:03 |
| aburgess | thank you, that looks helpful | 17:04 |
| poke53282 | Damn, it looks like someone broke the QEMU openrisc emulation during the last year. | 21:54 |
| stekern | some generic change that broke the target code? | 22:01 |
| poke53282 | I am investigating | 22:07 |
| poke53282 | Might be also my poor skills in using git. | 22:08 |
| stekern | gcc testing status, investigating this now: http://pastie.org/9848975 | 22:13 |
| poke53282 | looks ugly. | 22:16 |
| stekern | it fails in alpine as well | 22:16 |
| stekern | all my debug prints hasn't made it prettier ;) | 22:16 |
| poke53282 | Maybe also responsible for the one failed test in libffi | 22:16 |
| stekern | does it do unwind? | 22:16 |
| poke53282 | well, I had one problem with a ffi call and a try catch statement. between these calls. | 22:17 |
| stekern | it fails slightly different when I compile it natively in jor1k too | 22:17 |
| poke53282 | And try catch is an unwind. | 22:17 |
| stekern | and if I try to cross-compile it without static, it doesn't even invoke the signalhandler (fn5) | 22:19 |
| stekern | ...fn4 | 22:19 |
| stekern | (which calls fn5) | 22:20 |
| stekern | time to sleep on that | 22:20 |
| poke53282 | hehe, good night | 22:20 |
| --- Log closed Thu Jan 22 00:00:07 2015 | ||
Generated by irclog2html.py 2.15.2 by Marius Gedminas - find it at mg.pov.lt!