| # 00:57:09 | nandub | joins #crosstool-ng |
| # 02:07:31 | nandub | quits : Quit: Lost terminal |
| # 02:08:27 | nandub | joins #crosstool-ng |
| # 03:27:36 | jevin__ | joins #crosstool-ng |
| # 03:30:41 | jevin | quits : *.net *.split |
| # 03:30:42 | blueness | quits : *.net *.split |
| # 03:31:08 | jevin__ | is now known as: jevin |
| # 03:37:55 | blueness | joins #crosstool-ng |
| # 03:37:56 | blueness | quits : Changing host |
| # 03:37:56 | blueness | joins #crosstool-ng |
| # 05:38:28 | codyps | quits : Ping timeout: 248 seconds |
| # 05:50:53 | imMute | quits : Ping timeout: 255 seconds |
| # 05:51:05 | ubergeek42 | quits : *.net *.split |
| # 05:52:00 | rfs_ | joins #crosstool-ng |
| # 05:52:00 | imMute^ | joins #crosstool-ng |
| # 05:52:01 | imMute^ | quits : Changing host |
| # 05:52:01 | imMute^ | joins #crosstool-ng |
| # 05:52:35 | rfs | quits : Ping timeout: 260 seconds |
| # 05:56:47 | ubergeek42 | joins #crosstool-ng |
| # 06:25:25 | blueness | quits : Ping timeout: 248 seconds |
| # 06:26:04 | blueness | joins #crosstool-ng |
| # 06:26:04 | blueness | quits : Changing host |
| # 06:26:04 | blueness | joins #crosstool-ng |
| # 07:04:47 | sfan5|OFF | is now known as: sfan5 |
| # 08:52:38 | smartin | joins #crosstool-ng |
| # 09:26:01 | codyps | joins #crosstool-ng |
| # 10:28:14 | y_morin | joins #crosstool-ng |
| # 10:43:10 | y_morin | quits : Remote host closed the connection |
| # 10:57:49 | y_morin | joins #crosstool-ng |
| # 13:56:16 | y_morin_ | joins #crosstool-ng |
| # 14:01:27 | y_morin | quits : *.net *.split |
| # 14:01:36 | y_morin_ | is now known as: y_morin |
| # 14:12:19 | mwpowellhtx | joins #crosstool-ng |
| # 14:12:55 | mwpowellhtx | q: the latest crosstool release is up to date with GCC 4.7.2? |
| # 14:14:59 | nandub | quits : Quit: leaving |
| # 14:15:52 | y_morin | mwpowellhtx: yes. gcc-4.7.2 is available in 1.18.0 |
| # 14:16:46 | rfs_ | is now known as: rfs |
| # 14:17:14 | rfs | Hi all |
| # 14:18:20 | mwpowellhtx | so to turn on 0x support, we specify -std=c++0x or -std=c++11 for 11 ? |
| # 14:19:22 | y_morin | mwpowellhtx: yes, but this is done when using the cross-compiler (ie. at run-time, not at build-time). |
| # 14:20:10 | mwpowellhtx | run time of the cross compiler; i.e. build time of the app targeting the ARM? |
| # 14:21:06 | y_morin | mwpowellhtx: yes. |
| # 14:21:43 | y_morin | mwpowellhtx: I meant: not at build-time of the toolchain, but at run-time of the toolchain, which is at build-time of your app. |
| # 14:21:49 | y_morin | Confusing, isn't it? ;-) |
| # 14:21:53 | mwpowellhtx | not really |
| # 14:22:09 | y_morin | mwpowellhtx: Good, then! :-) |
| # 14:22:11 | mwpowellhtx | I've worked with VxWorks cross compilers, so some take a binary and convert that |
| # 14:22:22 | mwpowellhtx | this one actually builds source to the target |
| # 14:23:14 | mwpowellhtx | excellent news. any ideas when GCC 4.8 support? that will have additional concurrency I'll be interested in. |
| # 14:24:45 | blueness | quits : Ping timeout: 260 seconds |
| # 14:24:51 | mwpowellhtx | to verify type gcc --version and you get not only GCC version but also tool chain version |
| # 14:24:52 | y_morin | mwpowellhtx: when it will be released, which is expected not before end of March: http://gcc.gnu.org/ml/gcc/2013-02/msg00177.html |
| # 14:27:38 | mwpowellhtx | in the mean time, fair to say stick with 0x? or is 11 stable enough with the features it supports? |
| # 14:27:52 | y_morin | Well, 4.8.0-rc1 is expected in March. I'd suspect 4.8.0 final to be in April. |
| # 14:28:07 | y_morin | mwpowellhtx: I don't know. |
| # 14:28:33 | mwpowellhtx | thank ye |
| # 14:29:13 | y_morin | rfs: Hello! Any problem? |
| # 14:29:24 | Joeboy | parts #crosstool-ng |
| # 14:30:23 | rfs | y_morin: Morning :) Yep, I'm having trouble with glibc... compile fails in sem_post.c, seems to be that lll_futex_wake() no longer returns a value |
| # 14:30:56 | y_morin | rfs: Can you post your .config and your build.log? |
| # 14:31:26 | rfs | y_morin: sure, one sec. |
| # 14:31:57 | y_morin | rfs: not on the channel, use some pastebin for that. |
| # 14:35:12 | rfs | y_morin: http://nw.dreamhosters.com/ct-ng/ |
| # 14:36:30 | y_morin | rfs: that's because NPTL does not work on i386, which is too old a processor. You need at least i486 for NPTL. |
| # 14:37:17 | rfs | y_morin: I was actally aiming for i686... eg. 32-bit |
| # 14:38:26 | y_morin | rfs: then you need to set either or both of: CT_ARCH_TUNE or CT_ARCH_CPU |
| # 14:38:53 | y_morin | rfs: eg: CT_ARCH_TUNE="i686" |
| # 14:39:32 | y_morin | rfs: it's in the "Target options" sub-menu |
| # 14:41:55 | rfs_ | joins #crosstool-ng |
| # 14:42:29 | smartin | quits : Ping timeout: 245 seconds |
| # 14:42:29 | rfs | quits : Ping timeout: 245 seconds |
| # 14:43:13 | rfs_ | y_morin: so it defaults i386 when not otherwise specified? |
| # 14:43:17 | rfs_ | is now known as: rfs |
| # 14:43:33 | y_morin | rfs: yes |
| # 14:43:49 | y_morin | rfs: the default is gcc's choice, not crosstool-NG's |
| # 14:44:07 | rfs | y_morin: gotcha. Trying it now... |
| # 14:44:07 | smartin | joins #crosstool-ng |
| # 14:45:27 | rfs | y_morin: out of curiosity, how do you test various versions? I gather you use VMs, any particular tricks to the setup? |
| # 14:46:57 | y_morin | rfs: I mostly (95%) test on my Debian Squeeze (aka stable). I sometime fire up a VM with a different distro just to check a complex corner case that does not show up on my system. |
| # 14:47:16 | y_morin | rfs: otherwise, I rely on others to report issues and fixes for other distros. |
| # 14:48:19 | y_morin | rfs: With Debian, it's dead easy to build a chrrot with a different version, so I also sometime debootstrap a testing or an Ubuntu to check in the chroot rather than pay the overhead of a VM. |
| # 14:48:29 | y_morin | s/chrrot/chroot/ |
| # 14:49:03 | mwpowellhtx | quits : Quit: If you think nobody cares, try missing a few payments |
| # 14:51:28 | rfs | y_morin: ok thanks, I've been using chroot to test as well, but wasn't sure if that might be part of my problem. |
| # 14:52:24 | y_morin | rfs: rpobably not. Except if you are trying in a 32-bit chroot on an otherwise 64-bit machine. |
| # 14:54:11 | y_morin | rfs: but this was clearly not your issue. |
| # 14:54:12 | rfs | y_morin: yup, I've got all the 32-bit libs etc installed too. |
| # 14:54:12 | rfs | y_morin: I'm actually aiming for a self-building system, with just enough to be able to recompile itself. |
| # 14:54:12 | y_morin | rfs: no, I meant: a chroot with only the 32-bit libs, runing on a 64-bit kernel. |
| # 14:54:44 | rfs | y_morin: that's what I put in my chroot. All good. |
| # 14:54:56 | y_morin | rfs: and your kernel is 64-bit? |
| # 14:55:20 | rfs | y_morin: BTW, I put both CT_ARCH_TUNE and CT_ARCH_CPU and i'm getting warnings about -mcpu= is deprecated (but it is building) |
| # 14:55:30 | rfs | y_morin: yes, running on 64-bit host. |
| # 14:56:29 | y_morin | rfs: yes, -mcpu is deprecated, but it's only a warning. Leave it empty next time, and only set: CT_ARCH_TUNE CT_ARCH_ARCH |
| # 14:56:59 | y_morin | rfs: sorry, I initially gave you the wrong variables. The above two are the correct ones. |
| # 14:57:20 | rfs | y_morin: will do. Build completed successfully just now. |
| # 15:00:16 | rfs | y_morin: Many thanks for the diagnostics... |
| # 15:11:19 | mnt_real_ | joins #crosstool-ng |
| # 15:12:36 | rfs | re-build with CT_ARCH_TUNE and CT_ARCH_ARCH also completed successfully. Now to rebuild my chroot... |
| # 15:13:28 | mnt_real | quits : Ping timeout: 245 seconds |
| # 15:16:57 | sh4rm4 | quits : Quit: sh4rm4 |
| # 15:18:35 | sh4rm4 | joins #crosstool-ng |
| # 15:57:24 | diorcety | quits : Quit: Leaving. |
| # 16:20:43 | sfan5 | is now known as: sfan5|OFF |
| # 16:23:54 | rfs | quits : Quit: reboot |
| # 16:41:38 | mwpowellhtx | joins #crosstool-ng |
| # 16:42:39 | mwpowellhtx | q: trying to build a simple hello main c++ app. I type: /path/to/g++ -Wall main.cpp -o hello |
| # 16:43:07 | mwpowellhtx | I get: g++: error trying to exec 'cc1plus': execvp: No such file or directory |
| # 16:43:48 | y_morin | mwpowellhtx: Can you paste the exact /path/to/g++ ? |
| # 16:44:57 | mwpowellhtx | sure: /opt/x-tools/TARGET/arm-unknown-linux-gnueabi/bin/g++ |
| # 16:45:08 | mwpowellhtx | it's entirely possible I need to tweak my .config and rebuild |
| # 16:45:31 | mwpowellhtx | the linaro g++ works just fine. "Hello, world!" |
| # 16:45:38 | y_morin | mwpowellhtx: That's not what you should use. You should use: /opt/x-tools/TARGET/bin/TARGET-g++ |
| # 16:46:02 | y_morin | (obviously, replace 'TARGET' with 'arm-unknown-linux-gnueabi' |
| # 16:46:28 | mwpowellhtx | ok. but use TARGET-g++ (would be better to identify the ARM probably). I gotcha. |
| # 16:46:29 | y_morin | mwpowellhtx: the g++ you point to is an internal tool, and should not be used manually. |
| # 16:47:39 | y_morin | mwpowellhtx: you can export PATH="${PATH}:/opt/x-tools/TARGET/bin" and then use "TARGET-g++" without the full path. |
| # 16:47:56 | y_morin | mwpowellhtx: and exporting the PATH will allow you to run: ./configure --host=TARGET |
| # 16:48:18 | y_morin | mwpowellhtx: as ./configure will automatically find TARGET-{gcc,g++,as,ar,ld,...} |
| # 16:48:25 | mwpowellhtx | viola |
| # 16:48:34 | mwpowellhtx | file hello-arm tells me ARM info |
| # 16:49:03 | mwpowellhtx | golden tip re: configure. thank you |
| # 16:49:07 | y_morin | mwpowellhtx: I guess you meant "voilà " |
| # 16:49:12 | mwpowellhtx | yessir thanks |
| # 16:49:25 | y_morin | ;-) |
| # 16:49:42 | y_morin | should know about voilà , being French! ;-) |
| # 16:54:09 | mwpowellhtx | we'll probably need to adjust our kernel versions and such. I am more confident this puts on track to target our ARM. |
| # 16:56:53 | mwpowellhtx | quits : Quit: There's nothing dirtier then a giant ball of oil |
| # 16:59:25 | uminded | joins #crosstool-ng |
| # 17:02:08 | uminded | I can compile a toolchain (-march=armv7-a -mtune=cortex-a8 -mthumb -mfpu=neon -mfloat-abi=hard) and have compiled working kernels with it but I need a native chain to run on that cortex-a8. What do I set in menuconfig?? |
| # 17:03:25 | y_morin | uminded: crosstool-NG is not currently capable of building a native toolchain. |
| # 17:04:53 | y_morin | uminded: crosstool-NG is initially a cross-compilation tool. Having a toolchain on the target is not really the initial goal of crosstool-NG, and although I'd like to have it, I'm not investing too much on that for now... |
| # 17:04:53 | uminded | oh. Well it did a great job for my cross compiling one so I had hoped... What do you reccomended I do? I have not been able to find a download for a native cortex-a8 one |
| # 17:05:45 | y_morin | uminded: Well, I do recopmmend to do all with cross-compilation. There's no way your ARM target will build code as fast as your PC. (at least, not in the foreseeable future). |
| # 17:07:07 | uminded | Some of these crazy 2.5ghz quad core cell phones might change that soon... But I am trying to compile linuxcnc and currently if you use --host or CC= during configure it has a bug and fails |
| # 17:07:58 | uminded | it works fine when compiling nativly though so I was joing going to build that one on the arm |
| # 17:09:50 | y_morin | Argh... Yet another package that does not use a traditional build-system... :-( |
| # 17:10:14 | y_morin | uminded: You'll get faster at fixing linuxcnc that trying to build a native toolchain. |
| # 17:11:11 | y_morin | Ah, it uses autotools, in fact. |
| # 17:11:26 | uminded | yup |
| # 17:12:47 | y_morin | is going for a break... |
| # 17:16:03 | mwpowellhtx | joins #crosstool-ng |
| # 17:21:22 | diorcety | joins #crosstool-ng |
| # 17:24:02 | mwpowellhtx | uh oh ./configure command not found |
| # 17:24:27 | mwpowellhtx | I recently upgraded GCC to 4.7.2, then build the crosstool tool chain |
| # 17:30:18 | mwpowellhtx | <-- Linux newbie. configure is a script that gets included with a code base to configure Makefile(s)? |
| # 17:39:48 | hrubi | quits : Ping timeout: 264 seconds |
| # 17:42:00 | hrubi | joins #crosstool-ng |
| # 18:11:15 | y_morin | mwpowellhtx: what package are you trying to build? |
| # 18:12:05 | mwpowellhtx | I am learning what configure means to our code base, configuring our make system and such |
| # 18:13:00 | y_morin | mwpowellhtx: yes, some packages include a script named 'configure'. 'configure' is part of the autotools buildsystem; it is very common, but not all packages have. Some packages will use another buildsystem (eg. cmake...) |
| # 18:13:39 | y_morin | mwpowellhtx: if you're trying to build your own code, then you should know how to specify the cross-compiling tools, no? ;-) |
| # 18:14:14 | mwpowellhtx | we'll have a mix for decoupling, testability, etc, SOLID approach |
| # 18:18:48 | diorcety1 | joins #crosstool-ng |
| # 18:21:35 | diorcety | quits : Ping timeout: 260 seconds |
| # 19:00:33 | mwpowellhtx | quits : Quit: Some folks are wise, and some otherwise. |
| # 19:43:14 | sfan5|OFF | is now known as: sfan5 |
| # 21:10:09 | sfan5 | is now known as: sfan5|OFF |
| # 21:46:52 | mnt_real_ | quits : Quit: Leaving |
| # 22:20:49 | blueness | joins #crosstool-ng |
| # 22:20:49 | blueness | quits : Changing host |
| # 22:20:49 | blueness | joins #crosstool-ng |
| # 22:37:40 | smartin | quits : Quit: leaving |
| # 23:04:03 | y_morin | quits : Quit: Nighty Night! |