| *** rdale has quit IRC | 00:07 | |
| *** rdale has joined #baserock | 00:14 | |
| *** Stanto has quit IRC | 02:07 | |
| *** Stanto has joined #baserock | 02:13 | |
| paulsherwood | hmmmm. i'm trying scaleways.com (which seems pretty nice, actually).... | 08:41 |
|---|---|---|
| paulsherwood | but failing to build stage1-gcc... | 08:41 |
| paulsherwood | http://paste.baserock.org/orutakokos | 08:41 |
| paulsherwood | and http://paste.baserock.org/akobidarup | 08:41 |
| *** edcragg has joined #baserock | 09:06 | |
| *** edcragg has quit IRC | 09:20 | |
| *** petefoth_ has joined #baserock | 10:07 | |
| *** petefoth has quit IRC | 10:08 | |
| *** petefoth_ is now known as petefoth | 10:08 | |
| persia | paulsherwood: I'd check to see if MORPH_ARCH is being set correctly. | 12:54 |
| persia | From the output, it looks like this is so, but the GMP configure is the bit being modified here, and I don't see any output that looks like ./configure is being run again after that modification. | 12:56 |
| persia | So, either something isn't logged, or I can't read, or something is wrong with the modification. | 12:56 |
| persia | Note that this would differ from ARMv8 architectures, as in those the detection seems to work without adjustments (for reasons I do not entirely understand) | 13:00 |
| rjek | Do you mean aarch64? | 13:02 |
| rjek | ARMv8 includes (optional) support for aarch32, varying only very slightly from ARMv7 | 13:02 |
| persia | At least when I was testing things, morph distinguished strongly enough between 32-bit ARMv7 and 32-bit ARMv8 that it couldn't reliably build one from the other. | 13:33 |
| persia | Things may have gotten better (I haven't been following all the patches in this area), but given the time that ybd started development, I would be surprised if the equivalent patches landed in ybd, even if they did land in morph. | 13:34 |
| persia | The core toolchain seems especially sensitive to this: using cached core toolchain, I was able to build ARMv7 stuff under linux32 on ARMv8, but I never managed to build the toolchain that way, even with patched util-linux. | 13:35 |
| persia | (but I was doing this in November/December: things might be better now) | 13:36 |
| rjek | tool chain build systems are awful | 14:26 |
| rjek | film at 11 | 14:26 |
| persia | heh | 14:43 |
| paulsherwood | that build is supposed to be armv7lhf i think | 17:14 |
| rjek | https://mobile.twitter.com/davecheney/status/616931340466786304 | 17:25 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!