01:10.13 | *** join/#uclibc xiangfu (~xiangfu@fidelio.qi-hardware.com) |
01:14.12 | *** join/#uclibc mnt_real (~mnt_real@bas1-montreal43-2925255969.dsl.bell.ca) |
01:48.10 | *** join/#uclibc firegurafiku (~xf@45.160.219.91.infosys.su) |
03:22.04 | *** join/#uclibc shedammit (~shedammit@173-228-38-188.dsl.dynamic.sonic.net) |
03:22.04 | *** join/#uclibc shedammit (~shedammit@unaffiliated/shedammit) |
04:34.15 | *** join/#uclibc firegurafiku (~xf@22.160.219.91.infosys.su) |
05:04.51 | *** join/#uclibc fabled (~fabled@83.145.235.194) |
05:26.40 | *** join/#uclibc trem (~trem@bny93-7-88-161-33-221.fbx.proxad.net) |
05:50.28 | *** join/#uclibc ncopa (~ncopa@3.203.202.84.customer.cdi.no) |
06:50.27 | *** join/#uclibc smartin (52e7e007@gateway/web/freenode/ip.82.231.224.7) |
07:19.18 | *** join/#uclibc y_morin (c1317c6b@gateway/web/freenode/ip.193.49.124.107) |
07:23.39 | *** join/#uclibc sh4rm4 (~sh4rm@xdsl-188-155-204-78.adslplus.ch) |
07:25.03 | *** join/#uclibc acoul (~alex@orion.ozo.com) |
07:27.10 | *** join/#uclibc kos_tom (~thomas@col31-4-88-188-83-94.fbx.proxad.net) |
07:41.38 | CIA-79 | 03cernekee 07master * r2d4243ce60bc 10uClibc/ldso/ldso/dl-startup.c: ldso/mips: Enable bootstrap relocations |
07:41.38 | CIA-79 | 03cernekee 07master * ra2f827c7c28c 10uClibc/ldso/ldso/ (dl-hash.c mips/dl-sysdep.h): ldso/mips: dlsym() incorrectly matches undefined symbols |
07:41.38 | CIA-79 | 03cernekee 07master * r966adfe8ed03 10uClibc/ldso/ldso/mips/dl-sysdep.h: ldso/mips: Clean up warnings |
07:43.37 | CIA-79 | 03yegor_sub1 07master * r6378cac4e633 10buildroot/package/util-linux/ (7 files): util-linux: convert to autotarget and bump to 2.19.1 |
07:52.30 | kos_tom | Jacmet, yeah, yet another remaining package converted to the *TARGETS infrastructure \o/ |
07:56.09 | Jacmet | kos_tom: indeed |
07:57.11 | Jacmet | I should build a bunch of external toolchains for the different options we have (largefile, char, locale, ..) - It takes ages to test all those combinations from scratch on my old laptop |
07:59.01 | Jacmet | kos_tom: I'm thinking of getting rid of PROGRAM_INVOCATION. It only adds trivial overhead to uclibc (I would guess <100 bytes) and we have quite some packages needing it |
07:59.02 | *** join/#uclibc sh4rm4 (~sh4rm@unaffiliated/sh4rm4) |
08:10.08 | kos_tom | Jacmet, if the overhead is only <100 bytes, then go for it. |
08:11.27 | Jacmet | kos_tom: it is. It's just a question about keeping track of 2 pointers based on argv[0] in the uclibc startup code |
08:11.56 | Jacmet | kos_tom: only problem is that we then wouldn't catch issues with people using external uclibc based toolchains without this option enabled |
08:12.15 | Jacmet | kos_tom: but ok, it didn't work before either |
08:13.49 | Jacmet | kos_tom: this is basically the added code: http://git.buildroot.net/uClibc/tree/libc/misc/internals/__uClibc_main.c#n410 |
08:14.57 | kos_tom | Jacmet, well, there are many, many, many custom uClibc configurations that we don't support. |
08:15.16 | kos_tom | Jacmet, in terms of external uClibc toolchains, I think we should only try to support those that are widely available. |
08:15.35 | kos_tom | Jacmet, for example, the one made available for Blackfin are well-known for this architecture, so it makes sense to worry about supporting them |
08:15.55 | kos_tom | Jacmet, but we can't support any random external toolchain with a custom uClibc configuration that one can be build with ct-ng |
08:16.20 | *** join/#uclibc smartin (52e7e007@gateway/web/freenode/ip.82.231.224.7) |
08:17.51 | Jacmet | true |
09:03.58 | *** join/#uclibc `antonio` (~antonio@nat-sta-smtc2.tvu.ac.uk) |
09:22.05 | *** join/#uclibc luneff (~yury@84.51.197.242) |
09:28.58 | *** join/#uclibc shedammit (~shedammit@unaffiliated/shedammit) |
09:30.12 | *** join/#uclibc clio (~andrej@85.159.109.222) |
09:35.49 | *** join/#uclibc shedammit (~shedammit@173-228-38-188.dsl.dynamic.sonic.net) |
09:35.49 | *** join/#uclibc shedammit (~shedammit@unaffiliated/shedammit) |
09:56.56 | oktalogic | every binary built by buildroot is stripped by default, right? |
09:58.26 | CIA-79 | 03jacmet 07master * r213bbe9a2281 10buildroot/package/util-linux/util-linux.mk: util-linux: misc fixes |
09:58.27 | CIA-79 | 03jacmet 07master * r4ee9d569f9ae 10buildroot/ (19 files in 11 dirs): uClibc: drop BR2_PROGRAM_INVOCATION option |
09:59.00 | Jacmet | oktalogic: yes, everything is stipped in output/target unless you enable BR2_STRIP_none |
10:00.21 | curfont | so, I am building stuff for my phone which is armv7l |
10:00.25 | curfont | what is that in the menuconfig |
10:02.55 | Jacmet | curfont: cortex A8/A9 |
10:09.40 | *** join/#uclibc NIN101 (~NIN101@p5DD285D3.dip0.t-ipconnect.de) |
10:14.56 | curfont | thanks Jacmet |
10:15.12 | curfont | Jacmet: Any specific one or does it depend on the phone model? |
10:15.43 | curfont | SGS2 is Cortex A9 |
10:15.45 | curfont | ah |
10:22.47 | curfont | make seems to have errored |
10:34.56 | curfont | Unknown CPU used in --with-tune=cortex-a9 |
10:34.56 | curfont | make[1]: *** [configure-gcc] Error 1 |
10:37.22 | *** join/#uclibc luneff (~yury@109.165.54.167) |
10:43.05 | Jacmet | curfont: what gcc version did you select? |
10:43.22 | Jacmet | curfont: under toolchain options |
10:45.42 | curfont | Jacmet: i thought it was 4.4, seems I was wrong |
10:45.47 | curfont | changed it, lets see if it works |
10:46.08 | Jacmet | curfont: remember to do a make clean; make whenever you change toolchain options |
10:46.17 | curfont | aha, ok |
10:46.32 | curfont | thanks |
10:47.06 | *** join/#uclibc botao_sun (~botao_sun@linaro/botao-sun) |
10:53.28 | *** join/#uclibc _NIN (~NIN101@p5DD2863B.dip0.t-ipconnect.de) |
11:00.16 | *** join/#uclibc dileks_ (~dileks@2001:4dd0:f889:0:205:4eff:fe46:b41f) |
11:38.17 | curfont | Jacmet: Seems its giving me the classic "word unexpected", isnt this the endianess? |
11:38.23 | curfont | Is it arm vs armeb? |
11:44.26 | Jacmet | curfont: please put the exact error message + last few lines of output on pastebin |
11:50.33 | *** join/#uclibc bitterman (~yury@109.165.54.167) |
11:53.05 | *** join/#uclibc tiny (~ivob@unaffiliated/tiny) |
12:06.58 | curfont | Jacmet: bash: ./app: Permission denied |
12:07.06 | curfont | or app: 1: Syntax error: word unexpected (expecting ")") |
12:07.09 | curfont | when (sh) |
12:07.21 | curfont | also ---xrwxr-x 1 system sdcard_r 46300 Jul 27 12:05 app |
12:08.13 | Jacmet | curfont: what exactly are you trying to do? |
12:08.22 | curfont | Jacmet: I compiled it, ELF 32-bit LSB executable, ARM, version 1 (SYSV), statically linked, not stripped |
12:08.26 | curfont | Moved it to phone |
12:08.39 | curfont | It wont run, I had the same problem with I was doing MIPS and it had the wrong endianess |
12:09.08 | Jacmet | curfont: armeb is very rare, so that's unlikely to be the problem |
12:09.15 | curfont | The phone is: Linux localhost 2.6.35.7-g752b30b #1 SMP PREEMPT Thu Jul 21 14:43:46 EDT 2011 armv7l GNU/Linux |
12:09.26 | curfont | It is a Samsung Galaxy S 2 which is Cortex A9 |
12:09.46 | curfont | it is a simple hello world, to test it |
12:10.00 | Jacmet | curfont: anything of interest in your dmesg? |
12:12.58 | curfont | Nothing relevant, so for sure it is arm + Cortex A9? Maybe I need to do some other config? |
12:13.03 | curfont | I am using gcc 4.4 |
12:13.17 | kos_tom | curfont, did you link statically ? |
12:13.26 | curfont | yes: ELF 32-bit LSB executable, ARM, version 1 (SYSV), statically linked, not stripped |
12:13.31 | curfont | Thats what file shows ^ |
12:13.38 | kos_tom | ah, yes, sorry, didn't read the backlog properly. |
12:14.00 | kos_tom | the error message reminds me of something, but I don't remember what. |
12:14.36 | curfont | kos_tom: I got the same error message when I tried running a mips app with the wrong endianess |
12:15.13 | Jacmet | curfont: what does file say for one of the "native" programs on your phone? |
12:15.33 | Jacmet | curfont: perhaps its using thumb2 |
12:17.07 | curfont | cat: ELF 32-bit LSB executable, ARM, version 1 (SYSV), dynamically linked (uses shared libs), stripped |
12:17.26 | curfont | strange |
12:18.31 | Jacmet | curfont: no idea then |
12:19.06 | curfont | Is android phone dev usually done using uclibc? |
12:20.18 | Jacmet | curfont: no |
12:20.36 | Jacmet | curfont: google has its own libc |
12:20.58 | curfont | brilliant Jacmet, it worked |
12:21.05 | curfont | "stripping" |
12:21.19 | curfont | that shouldnt even cause this problem right? |
12:21.34 | curfont | seems android phones dont like unstripped executables |
12:26.03 | oktalogic | hm, stripping should only remove debug info and such no..? not affect the execetion otherwise..? |
12:27.32 | curfont | thats what I thought too |
12:27.46 | curfont | unless wget didn't get the file properly the first time |
12:27.48 | curfont | oh well |
12:32.26 | curfont | Is there some kind of script in the buildroot that will set the environment so make uses the correct gcc? |
12:35.05 | kos_tom | curfont, no we don't have anything like this at the moment |
12:38.19 | Jacmet | curfont: would that be anything else than export PATH=$PATH:$PWD/output/host/usr/bin ? |
12:38.33 | *** join/#uclibc Artemys (~quassel@stp25-2-82-234-232-91.fbx.proxad.net) |
12:38.48 | kos_tom | Jacmet, we could have a helper that sets an environment with CC=blabla-gcc LD=blabla-ld, etc. |
12:39.19 | curfont | Jacmet: usually, binaries are named arm-linux-gcc |
12:39.35 | curfont | Aha, I could make symlinks |
12:39.40 | curfont | and then set the path to them |
12:39.49 | Jacmet | kos_tom: true, we could. Wouldn't that cause more problems than it would fix for people who also wants to compile something with their system compiler now and then? |
12:39.59 | curfont | Hmm |
12:44.04 | kos_tom | Jacmet, we can document something like "Run commands this way: (source ~/buildroot/output/br-env.sh ; make)" |
12:44.25 | kos_tom | but it's a bit horrible |
12:45.05 | kos_tom | or make $(cat ~/buildroot/output/br-env.sh) |
12:45.39 | kos_tom | or simply, we could document what values should have the different variables |
12:46.17 | kos_tom | Jacmet, generally speaking, I'd like to improve our Buildroot documentation, add some tutorials, and stuff like that. I think we have reached a relatively good level of quality/stability (for the existing feature set), but we lack docs. |
12:46.36 | curfont | buildroot is awesome and simple if you just use gcc on its own |
12:46.43 | curfont | But makefiles, autoconf etc are a nightmare |
12:46.52 | Jacmet | kos_tom: yes. I don't really feel strongly about it. Here, I just tell people to add host/usr/bin to the end of the path and people know how to pass CROSS_COMPILE to make |
12:51.51 | curfont | has anyone built things like ncurses with uclibc |
12:52.21 | kos_tom | curfont, it's available as a package in BR, so yes. |
12:52.43 | curfont | ah, so there is a way to get it |
12:54.47 | curfont | kos_tom: http://buildroot.uclibc.org/buildroot.html#add_packages |
12:54.49 | curfont | is this how I add it? |
12:58.25 | kos_tom | curfont, no, ncurses is already available as a package, so someone has already done this work. Just run menuconfig and tick the option. |
14:09.29 | *** join/#uclibc mnt_real (~mnt_real@bas1-montreal43-2925256586.dsl.bell.ca) |
14:19.05 | *** join/#uclibc kos_tom (~thomas@col31-4-88-188-83-94.fbx.proxad.net) |
14:24.15 | *** join/#uclibc n17ikh (~n17ikh@c-174-56-150-44.hsd1.sc.comcast.net) |
14:27.37 | *** join/#uclibc antgreen (~user@bas3-toronto06-1177894329.dsl.bell.ca) |
14:38.15 | *** join/#uclibc julm (~julm@195.88.84.18) |
15:00.05 | *** join/#uclibc pedwo (3e49aa13@gateway/web/freenode/ip.62.73.170.19) |
15:02.19 | pedwo | Hi, can anyone point me to a good example of a package that builds host and target programs in one pass? Or do we have to build once for host and then again for the target? |
15:33.37 | *** join/#uclibc PasNox (~pasnox@2a01:e35:8a99:6cd0:6ef0:49ff:fe58:2614) |
15:39.39 | *** join/#uclibc fabled (~fabled@83.145.235.194) |
15:51.14 | *** join/#uclibc d33tah (~d33tah@are158.neoplus.adsl.tpnet.pl) |
15:51.23 | quotrem | pedwo: since they (usually) use different compilers/toolchains there's no way except to do separate host and target builds. |
15:51.29 | d33tah | hi guys, i've got a busybox issue |
15:52.02 | kos_tom | quotrem, well it depends. There are some tools that understand the difference between CC and HOSTCC |
15:52.08 | d33tah | when I compile it with a single applet, it complains "ether" is undeclared (i compiled it to support only ether-wake) |
15:52.10 | kos_tom | where CC = HOSTCC in native build, but CC != HOSTCC in cross-builds |
15:53.14 | quotrem | kos_tom: yeah but you have to do two builds right? "CC=host-gcc make xxx; CC=target-gcc make xxx" or equivalent, right? |
15:53.48 | d33tah | it's the latest busybox and i feel kinda lost |
15:54.32 | pedwo | quotrem, kos_tom: looking at it a bit more the host tool depends on a lib that is also built for the target so it will have to be seperate passes |
15:54.50 | pedwo | this is in the latest version of DirectFB... |
16:06.02 | *** join/#uclibc linuxjacques (6217fec1@nslu2-linux/jacques) |
16:16.36 | *** join/#uclibc risca (~risca@m90-129-51-104.cust.tele2.se) |
16:24.06 | d33tah | okay, i know what the problem is |
16:25.07 | d33tah | you can't compile single-applet ether-wake most likely because the 'ether-wake' name contains a hyphen |
16:25.18 | d33tah | i renamed it to etherwake and it works awesome. |
16:46.16 | *** join/#uclibc dileks_ (~dileks@2001:4dd0:f889:0:205:4eff:fe46:b41f) |
16:51.55 | *** join/#uclibc botao_sun (~botao_sun@linaro/botao-sun) |
17:00.04 | d33tah | https://bugs.busybox.net/show_bug.cgi?id=4009 |
17:08.43 | *** join/#uclibc risca (~risca@m90-129-47-5.cust.tele2.se) |
17:14.02 | *** join/#uclibc Jazzinghen (~jazzinghe@host15-192-dynamic.14-87-r.retail.telecomitalia.it) |
17:14.29 | Jazzinghen | Hi |
17:17.39 | Jazzinghen | I just passed by to tell y_morin and kos_tom that I've solved (in a way) the problem that I had yesterday, but I notice that the are not here... |
17:19.14 | *** join/#uclibc kos_tom (~thomas@humanoidz.org) |
17:21.12 | Jazzinghen | kos_tom: oh, hey! I was just talking about you |
17:21.27 | Jazzinghen | kos_tom: tank you again for you help yesterday |
17:22.34 | Jazzinghen | kos_tom: I was able to fix some things up, in the end. Today I managed, using crosstool and hacking my way through it, to compile a toolchain exactly for Kernel 2.4.20 with glibc 2.3.2 |
17:22.56 | Jazzinghen | kos_tom: it was Hell, but I managed to compile and run a dropbear server |
17:23.49 | Jazzinghen | kos_tom: the gcc I've compiled for that platform is even more recent than the one I use on my PC. |
17:25.15 | kos_tom | good! |
17:57.54 | *** join/#uclibc luneff (~yury@84.51.199.95) |
18:09.02 | *** join/#uclibc bitterman (~yury@84.51.199.95) |
18:54.40 | *** join/#uclibc firegurafiku (~xf@79.160.219.91.infosys.su) |
19:17.21 | *** join/#uclibc smartin (52e20d9c@gateway/web/freenode/ip.82.226.13.156) |
19:36.59 | *** join/#uclibc risca (~risca@m90-129-166-75.cust.tele2.se) |
19:49.02 | *** join/#uclibc d33tah (~d33tah@are158.neoplus.adsl.tpnet.pl) |
19:54.58 | CIA-79 | 03yegor_sub1 07master * r1f2879ccb7eb 10buildroot/package/usbmount/ (4 files): usbmount fixes |
19:54.59 | CIA-79 | 03jacmet 07master * r4e3ae5ccaba2 10buildroot/package/busybox/Config.in: busybox: BR2_PACKAGE_BUSYBOX_SHOW_OTHER: remove conditional check |
20:02.35 | *** join/#uclibc shedammit (~shedammit@unaffiliated/shedammit) |
20:07.21 | *** join/#uclibc washer (~washer@75-141-232-86.dhcp.reno.nv.charter.com) |
20:35.18 | firegurafiku | hm.... make: *** No rule to make target `linux26', needed by `/arc/projects/tmp/buildroot/output/images/rootfs.iso9660'. Stop |
20:35.18 | firegurafiku | wtf? (( |
20:35.56 | *** join/#uclibc firegurafiku (~xf@79.160.219.91.infosys.su) |
20:41.44 | smartin | hi all, does anyone know what are the default Qt libs? |
20:43.40 | smartin | <PROTECTED> |
21:05.24 | *** join/#uclibc xf_ (~xf@61.160.219.91.infosys.su) |
21:07.00 | *** join/#uclibc y_morin (~ymorin@ARennes-256-1-84-132.w90-32.abo.wanadoo.fr) |
21:27.17 | CIA-79 | 03jacmet 07master * r3612ecf4a86f 10buildroot/fs/iso9660/iso9660.mk: iso9660: fix build after linux package change |
21:41.05 | *** join/#uclibc trem (~trem@bny93-7-88-161-33-221.fbx.proxad.net) |
21:58.53 | *** join/#uclibc xf_ (~xf@10.160.219.91.infosys.su) |
22:02.17 | CIA-79 | 03jacmet 07master * ra4a20acb2e7a 10buildroot/CHANGES: CHANGES: update with recent changes |
22:08.16 | y_morin | Jacmet: hey ho! :-) |
22:09.14 | y_morin | Jacmet: two quickies : are you going VIP? Did you have time to complete the CFP mail? |