ibotlog2html for #crosstool-ng

<< Previous 2011-08-31 Next >>

# 00:16:13 xvilka quits : *.net *.split
# 00:20:22 mnt_real joins #crosstool-ng
# 00:21:01 xvilka joins #crosstool-ng
# 00:21:01 xvilka quits : Changing host
# 00:21:01 xvilka joins #crosstool-ng
# 00:22:51 Buglouse quits : Remote host closed the connection
# 00:29:06 Buglouse joins #crosstool-ng
# 02:30:56 Tenacious-Techhu Anyone here know if crosstool-ng works under OS X yet?
# 02:46:10 Tenacious-Techhu quits : Quit: Page closed
# 04:37:13 mnt_real quits : Remote host closed the connection
# 04:45:12 bhundven|afk is now known as: bhundven
# 05:12:27 tor joins #crosstool-ng
# 06:40:48 kos_tom joins #crosstool-ng
# 06:53:50 y_morin joins #crosstool-ng
# 07:18:17 Buglouse quits : Read error: Operation timed out
# 07:22:53 y_morin is now known as: y_morin|away
# 07:33:45 y_morin|away is now known as: y_morin
# 08:21:42 bhundven is now known as: bhundven|afk
# 08:24:40 rogutes parts #crosstool-ng
# 08:29:41 groton joins #crosstool-ng
# 08:32:05 y_morin is now known as: y_morin|away
# 08:34:27 y_morin|away is now known as: y_morin
# 09:08:41 groton quits : Quit: Leaving.
# 09:09:44 groton joins #crosstool-ng
# 09:09:48 groton quits : Changing host
# 09:09:48 groton joins #crosstool-ng
# 10:10:23 y_morin is now known as: y_morin|away
# 11:00:50 y_morin|away is now known as: y_morin
# 11:28:52 groton quits : Quit: Leaving.
# 12:00:33 mnt_real joins #crosstool-ng
# 12:01:25 y_morin is now known as: y_morin|away
# 12:10:18 y_morin|away is now known as: y_morin
# 12:22:53 y_morin is now known as: y_morin|away
# 12:35:06 groton joins #crosstool-ng
# 13:01:21 mnt_real quits : Remote host closed the connection
# 13:04:29 mnt_real joins #crosstool-ng
# 13:26:23 y_morin|away is now known as: y_morin
# 13:47:23 y_morin is now known as: y_morin|away
# 13:48:56 y_morin|away is now known as: y_morin
# 14:48:41 Buglouse joins #crosstool-ng
# 14:58:12 y_morin is now known as: y_morin|away
# 15:07:02 y_morin|away is now known as: y_morin
# 15:11:55 Buglouse quits : Ping timeout: 258 seconds
# 15:27:06 Buglouse joins #crosstool-ng
# 15:36:50 Buglouse quits : Ping timeout: 258 seconds
# 15:39:04 Buglouse joins #crosstool-ng
# 15:50:02 y_morin is now known as: y_morin|away
# 15:59:18 y_morin|away is now known as: y_morin
# 16:41:00 y_morin is now known as: y_morin|away
# 16:52:10 kos_tom quits : Quit: Leaving
# 17:35:46 kos_tom joins #crosstool-ng
# 17:59:34 snappy y_morin|away: wen you get back, mind assisting me. When I'm compiling using armv7 as target cpu, libgcc fails and says "cc1: target CPU does not support ARM mode"
# 18:13:57 groton quits : Quit: Leaving.
# 18:25:24 y_morin|away is now known as: y_morin
# 18:36:43 y_morin is now known as: y_morin|away
# 18:49:47 groton joins #crosstool-ng
# 18:49:48 groton quits : Changing host
# 18:49:48 groton joins #crosstool-ng
# 19:04:41 mnt_real quits : Remote host closed the connection
# 19:06:26 mnt_real joins #crosstool-ng
# 20:04:38 groton quits : Quit: Leaving.
# 20:32:59 y_morin|away is now known as: y_morin
# 20:35:50 y_morin snappy: put your complete build.log file on-line, please.
# 20:36:05 snappy y_morin: sure
# 20:39:42 snappy It's a little large: http://lastninja.net/~naveen/ct/
# 20:39:50 snappy I'm using ct-ng 1.12.0
# 20:41:34 y_morin snappy: getting log file...
# 20:45:38 snappy i also put up the config.log that's referenced in the build.log just incase, but i'm not sure where the problem stems
# 20:45:48 snappy im starting to think versions of gcc/binutils/etc.
# 20:47:36 y_morin snappy: ah, config.log, that's what I was gonna ask for after reading the build.log
# 20:52:51 y_morin snappy: I am afraid gcc-4.4 is too old for Neon. Also, crosstool-NG does not (yet?) fully supports HW floating point such as Neon. There is work in this area (a very recent post on the list), but there is no guarantee at all. Pure soft float is what has been the most thoroughly tested so far.
# 20:54:07 y_morin snappy: also, you have a bunch of experimental options. You should first try without experimental, until you get a working combination. *Then* you can try adding experimental features, such as interworking, neon FPU, and so on...
# 20:54:20 y_morin snappy: I'll try here with more recent versions...
# 20:55:12 snappy ah ok; i was using configs supplied by xvilka, but tweaekd them a little (he was using it for OMAP)
# 20:55:27 snappy but no i have some stuff to try, better than hitting my head against the wall
# 20:56:28 y_morin snappy: it's easier to start from a working state, and then iterate one thing at a time. Yes, it is time consuming. Yes, having a fast PC or a good espresso machine helps! :-)
# 20:57:44 snappy so i think this machine im working on actually has 8 cores, but i haven't checked if ct-ng has any support for parallelism (or invoking make -j *when it works*)
# 20:58:42 y_morin snappy: either use the menuconfig option CT_PARALLEL_JOBS, or use: ct-ng build.NB_PROCS (eg. for make -j12: ct-ng build.12)
# 20:59:00 snappy ah nice
# 20:59:11 y_morin snappy: I always use NB_PROCS = 1.5*NB_CPUS
# 20:59:24 y_morin snappy: so I use: ct-ng build.18
# 20:59:49 snappy oh that's awesome.
# 21:04:20 tor quits : Quit: Ex-Chat
# 21:40:56 y_morin snappy: 'armv7' is wrong. The real value is 'armv-a'
# 21:41:07 y_morin snappy: but then it breaks in uClibc
# 21:42:28 y_morin snappy: OK, found the uClibc error. 'gold' is not ready for prime-time, it seems... :-(
# 21:46:13 snappy oh armv-a? hm/
# 21:46:24 snappy damn, forgot that cpu name again
# 21:46:42 y_morin snappy: ouch... 'armv7-a'... v_7_!
# 21:46:55 snappy one sec, let me link you to a site on arm
# 21:47:02 snappy in oticed that ARMV7 != ARMV7-A
# 21:48:03 y_morin snappy: I am only speaking gcc values here. I don't know/care how ARM calls these, but gcc seems to imply that 'armv7' is thumb only, while armv7-a is arm
# 21:48:20 snappy hm, i just found something interesting on android website though
# 21:48:22 snappy ARMv5TE machine code will run on all ARM-based Android devices. ARMv7-A will run only on devices such as the Verizon Droid or Google Nexus One that have a compatible CPU. The main difference between the two instruction sets is that ARMv7-A supports hardware FPU, Thumb-2, and NEON instructions. You can target either or both of the instruction sets — ARMv5TE is the default, but switching to ARMv7-A is as easy as adding a single line to the applica
# 21:48:29 snappy (sorry if there's a cutoff)
# 21:48:30 snappy http://developer.android.com/sdk/ndk/overview.html
# 21:48:36 snappy basically i could target ARMv5TE it seems like
# 21:49:47 y_morin snappy: best bet is to start off with armv5te, indeed. But if your real processor is better, then you could gradually upgrade to armv7-a...
# 21:50:02 snappy sounds like a plan, i tihnk i'll try this for now
# 21:50:16 y_morin snappy: I'm reaading on the floating point modes on ARM... What a messs... http://wiki.debian.org/ArmHardFloatPort/VfpComparison
# 21:50:51 snappy heh wow, look spretty bad
# 21:51:03 y_morin snappy: "The combination of -mfpu=vfp and -mfloat-abi=hard is not available in FSF GCC 4.4, but the CodeSourcery 2009q1 compiler (and later, current release is 2010q1) supports it as does FSF GCC 4.5. "
# 21:51:47 y_morin snappy: ermmm. I somewhat interpreted this as 'neon' in place of 'vfp'...
# 21:52:18 y_morin but I remember that neon also requires gcc-4.5+
# 21:55:39 Tenacious-Techhu joins #crosstool-ng
# 21:55:49 Tenacious-Techhu Hello. Anyone here today?
# 21:56:28 Tenacious-Techhu Can anyone tell me about the status of crosstool-ng on OS X?
# 21:59:15 snappy y_morin: thanks, going to try a lot of fiddling around now
# 21:59:56 y_morin snappy: I have a .config that builds a cortex-a8 optimised toolchain. Will post...
# 22:00:41 y_morin snappy: http://code.bulix.org/yd982z-80497
# 22:02:09 kos_tom y_morin, hi!
# 22:02:19 snappy thanks
# 22:02:24 y_morin kos_tom: hello there! :-)
# 22:02:28 kos_tom y_morin, during my holidays, I have worked on sh2a support in CT-NG. I made some progress but still there are some issues.
# 22:02:34 y_morin kos_tom: how's it been
# 22:02:35 kos_tom y_morin, are you interested in the current status of the work ?
# 22:02:50 y_morin kos_tom: sure!
# 22:03:24 kos_tom ok, I initially didn't use mercurial queues. After a bit of fighting, I manage to generate patches for the work I did, but I don't know how to import those patches into a mercurial queue
# 22:03:53 y_morin kos_tom: hg qimport *.patch
# 22:04:54 y_morin kos_tom: hmmm better be name-ordered (eg. 000-blabla.patch 001-boo.patch...)
# 22:05:29 kos_tom y_morin, ok, will try, thanks!
# 22:05:45 y_morin kos_tom: cheers!
# 22:09:51 xvilka quits : Read error: Operation timed out
# 22:10:59 xvilka joins #crosstool-ng
# 22:10:59 xvilka quits : Changing host
# 22:10:59 xvilka joins #crosstool-ng
# 22:50:35 groton joins #crosstool-ng
# 22:50:50 groton quits : Client Quit
# 22:51:26 Tenacious-Techhu Can someone please tell me about the status of crosstool-ng on OS X?
# 22:59:52 y_morin Tenacious-Techhu: 1.12.1 should be pretty OK.
# 23:00:17 y_morin Tenacious-Techhu: you /may/ get a few issues in some corner-cases, but all-in-all, it works.
# 23:03:44 Tenacious-Techhu Does it install easily? The main web page seems to suggest it has some issues.
# 23:03:52 Tenacious-Techhu Or is that out of date?
# 23:04:30 y_morin Tenacious-Techhu: the site is out-of-date. Badly, at that...
# 23:05:10 Tenacious-Techhu Is there a helpful install guide or something?
# 23:05:31 y_morin Tenacious-Techhu: http://crosstool-ng.org/#download_and_usage ;-)
# 23:24:29 y_morin quits : Quit: Time for rest...
# 23:57:26 Buglouse quits : Remote host closed the connection

Generated by ibotlog2html by Yann E. MORIN