2013-05-13T00:21:27 sorry kiwichris_ something came up, I had to leave for a bit. 2013-05-13T00:21:34 Anyways, heres the pastebin http://pastebin.com/crvcRrPK 2013-05-13T00:23:35 huh.. it pasted strangely. Blame cmd.exe ... let me try using another shell 2013-05-13T00:27:20 hmm.. can't find a better way to copy. Every shell on windows I got seems to copy the same way.. sorry about htat 2013-05-13T00:35:23 hi,the9a3eedi, does you hellomodule.c contains an "rtems" function? 2013-05-13T00:40:46 freenix, I've defined a task in that file, yes. I was hoping I could load tasks dynamically 2013-05-13T00:40:56 Is it a problem? 2013-05-13T00:42:12 I have not tried to creat tasks in the rap file. I saw you said the "rtems" symbol can not be found when loading. 2013-05-13T00:43:19 If "rtems" symbol can not be found when you load the rap file, you need to create a function named "rtems" in your source file, because the loader uses the "rtems" function as the entry. 2013-05-13T01:41:18 freenix, I think I understood you incorrectly. I do not have a function or any variable in my source named "rtems".. but I am using the rtems API to define a task (called "HelloTask") 2013-05-13T01:42:47 you should define a function named "rtems", and this will be used as the entry. 2013-05-13T01:43:20 just like this: int rtems(int argc, char **argv) {.....; return 0} 2013-05-13T01:44:48 what should the entry function do? I am under the impression that the rtems linker loads a bunch of functions and does not need an entry function 2013-05-13T01:45:15 I will try to make an "rtems" function anyway. Will see how that goes 2013-05-13T01:47:01 you may look insight into the file "rap.c" of the loader. line 344. 2013-05-13T02:08:20 I just took a look. Makes sense. Thanks freenix for the info 2013-05-13T02:09:25 :) 2013-05-13T02:10:11 freenix, thanks 2013-05-13T02:11:07 hi, kiwichris_. 2013-05-13T02:11:43 I just share what I know.:) 2013-05-13T02:17:10 :) 2013-05-13T02:22:39 *** sebhub has joined #rtems 2013-05-13T02:22:45 good morning 2013-05-13T02:23:47 moorning 2013-05-13T02:28:09 sebhub, hi 2013-05-13T02:28:46 I seem to remember something recently about a Unix port for RTEMS. I cannot find the link. 2013-05-13T02:29:31 more a hack than a port, i can send you the latest patch 2013-05-13T02:36:24 Yes please. 2013-05-13T02:36:32 What is missing ? 2013-05-13T02:37:48 i had no time to really look at it 2013-05-13T02:38:06 integration has currently the lowest priority for me 2013-05-13T02:40:46 ah ok 2013-05-13T02:48:06 defining an entry task in the source solves the error, but now I'm getting another error: 2013-05-13T02:48:14 error: rap::write: Image output size does not match layout size: .text: layout-size=0 image-size=116 2013-05-13T02:48:27 stumped again. Am I missing something? 2013-05-13T02:49:31 This is a bug. 2013-05-13T02:50:04 Please raise a PR with the code to recreate 2013-05-13T02:51:02 sebhub, with Ralf K's patch to gcc, any thoughts on handling this in the RSB ? I assume turning it on is ok if you do not use these encodings ? 2013-05-13T02:51:20 I mean gcc building 2013-05-13T02:52:00 ie is this a feature we should always have on or something a user may want ? 2013-05-13T02:52:50 this patch will add all (I hope) code pages to the Newlib iconv() 2013-05-13T02:52:56 if you don't use iconv() nothing changes 2013-05-13T02:53:13 So this should be on by fault ? 2013-05-13T02:53:18 if you use iconv() i wonder how you did this with the current version, since no code pages were present 2013-05-13T02:53:32 No idea. 2013-05-13T02:53:53 an alternative to the static approch is a dynamic, you can install code pages at run time 2013-05-13T02:54:08 but i have no idea how this works, i asked the newlib list and got no answer 2013-05-13T02:54:10 Add an sb-rtems-config to the RSB. It lets you ask a BSP for specific options, flags etc 2013-05-13T02:54:27 Add -> Added 2013-05-13T02:54:34 this iconv stuff is not bsp specific 2013-05-13T02:54:44 Sure. Sorry changed topic :) 2013-05-13T02:54:47 oh 2013-05-13T02:55:17 This command sb-rtems-config is a poor version of the intended rtems-config that uses the Makefile.inc hack to get the values 2013-05-13T02:55:47 It should allow integration of the addons into the RSB 2013-05-13T02:55:58 you are quite productive ;-) 2013-05-13T02:56:44 Hmm does that mean you think it is a waste of time ? :) 2013-05-13T02:57:22 you add new features faster than I can try them out 2013-05-13T02:57:49 It is to solve the problem Joel and others have had for years. If I did not I feared something else much more horrible would be added to the RSB that actually used the Makefile.incs 2013-05-13T02:59:58 verm__, I see a devel/mingw64-binutils but no devel/mingw64-gcc port ? Is there one ? 2013-05-13T03:00:36 kiwichris_, I'll raise a PR, but first I'll try other code samples to see if it's really the tool and not my code. 2013-05-13T03:00:51 kiwichris_: 64bit? i don't think so 2013-05-13T03:00:53 not yet, anyway 2013-05-13T03:01:09 the9a3eedi, please raise the PR no matter what. This is a bug and I need to hunt it down. 2013-05-13T03:01:19 is the rtems linker designed only to work with the git version (4.11)? because I'm working with 4.10.2 tools.. maybe that's why I'm getting issues? 2013-05-13T03:01:20 verm__, shame 2013-05-13T03:01:41 the9a3eedi, it is designed to work with ELF 2013-05-13T03:01:52 so in theory it should not matter 2013-05-13T03:02:00 kiwichris_, ok, will do. I might take a look at the code myself if I have the time 2013-05-13T03:02:30 Could you please xz compress the objects and attach them to save me rebuilding them 2013-05-13T03:05:46 Will do. 2013-05-13T03:06:10 More importantly, does this mean that the rtems-linker is not suited for production use? I was under the impression that it is, since the wiki said it's verified for sparc and all 2013-05-13T03:14:54 It is new code and while I would love to think it is it currently needs more work 2013-05-13T03:15:22 Unless users like you bang away using it it will sit so please keep using it 2013-05-13T03:18:05 ....... sorry... how do you raise a PR? I feel stupid for not asking earlier. I assumed a PR is another name for a "bug report" based on context 2013-05-13T03:18:39 www.rtems.org on the right hand side 2013-05-13T03:19:16 verm__, a few minor tweaks and it would seem RSB is building mingw32 exe's on FreeBSD :) 2013-05-13T03:19:26 need to check what it generates 2013-05-13T03:19:54 ah, bugzilla. 2013-05-13T03:20:28 awesome 2013-05-13T03:20:42 hi kiwichris_, the9a3eedi. 2013-05-13T03:20:52 I have paste a patch here:https://gist.github.com/MrVan/5566887 2013-05-13T03:21:16 hi, the9a3eedi. you may try the patch. Hope it can help resolve you problem 2013-05-13T03:21:20 that was quick :) I haven't even reported the bug yet 2013-05-13T03:21:33 will check it out 2013-05-13T03:23:55 *** Fallenou_ has joined #rtems 2013-05-13T03:24:44 *** noselasd_ has joined #rtems 2013-05-13T03:25:09 *** antgreen_ has joined #rtems 2013-05-13T03:27:03 *** rokka_ has joined #rtems 2013-05-13T03:27:12 *** budrys__ has quit IRC (*.net *.split) 2013-05-13T03:27:13 *** antgreen has quit IRC (*.net *.split) 2013-05-13T03:27:14 *** dhananja1 has quit IRC (*.net *.split) 2013-05-13T03:31:01 *** noselasd has quit IRC (*.net *.split) 2013-05-13T03:31:02 *** Fallenou has quit IRC (*.net *.split) 2013-05-13T03:31:04 *** rokka has quit IRC (*.net *.split) 2013-05-13T03:32:33 *** budrys__ has joined #rtems 2013-05-13T03:32:33 *** dhananja1 has joined #rtems 2013-05-13T03:36:34 *** budrys__ has quit IRC (*.net *.split) 2013-05-13T03:36:35 *** dhananja1 has quit IRC (*.net *.split) 2013-05-13T03:37:24 *** budrys__ has joined #rtems 2013-05-13T03:37:24 *** dhananja1 has joined #rtems 2013-05-13T04:03:47 freenix, the patch fixed the issue. Thanks! 2013-05-13T04:03:57 kiwichris_, should I still bother with the bug report? lol 2013-05-13T04:04:59 note: I haven't actually tried to load the rap file, so I can't say if it generated it correctly. The patch simply made the command work 2013-05-13T04:06:45 hi, the9a3eedi. I have tried it on ppc. I think it should also work on your platform. 2013-05-13T05:10:13 *** the9a3eedi has quit IRC (Ping timeout: 248 seconds) 2013-05-13T05:34:50 *** the9a3eedi has joined #rtems 2013-05-13T06:12:03 *** the9a3eedi has quit IRC (Read error: Connection reset by peer) 2013-05-13T06:12:07 *** the9a3eedi has joined #rtems 2013-05-13T06:12:13 *** Fallenou_ has quit IRC (Ping timeout: 260 seconds) 2013-05-13T06:15:02 *** Fallenou has joined #rtems 2013-05-13T06:19:43 *** MegaAlex|away is now known as MegaAlex 2013-05-13T06:21:29 *** the9a3eedi has quit IRC (*.net *.split) 2013-05-13T06:21:32 *** gedare has quit IRC (*.net *.split) 2013-05-13T06:25:22 *** the9a3eedi has joined #rtems 2013-05-13T06:25:22 *** gedare has joined #rtems 2013-05-13T07:04:46 *** MegaAlex has quit IRC (Ping timeout: 256 seconds) 2013-05-13T07:05:29 *** arvind_khadri has joined #rtems 2013-05-13T07:10:02 *** MegaAlex|away has joined #rtems 2013-05-13T07:34:38 *** weiY has joined #rtems 2013-05-13T07:43:46 *** antgreen_ has quit IRC (Quit: Leaving) 2013-05-13T07:50:36 *** noselasd_ has quit IRC (*.net *.split) 2013-05-13T07:51:04 *** the9a3eedi has quit IRC (Read error: Connection reset by peer) 2013-05-13T07:51:05 *** noselasd has joined #rtems 2013-05-13T07:51:31 Hi gedare. 2013-05-13T07:51:44 hey weiY 2013-05-13T07:51:54 *** the9a3eedi has joined #rtems 2013-05-13T07:51:58 i switch to RSB and the issue is disappared 2013-05-13T07:52:26 it seems the problem is toolchain package 2013-05-13T07:52:59 which problem? 2013-05-13T07:53:00 and now i should found a way to test mips SMP atomic ops 2013-05-13T07:53:44 when i use toolchain pulled from package, when i build mips architecture it will use stack check and it will report error 2013-05-13T07:55:00 with the rpm tools you get an error and with the RSB build one not? 2013-05-13T07:55:29 yeah, wait. and i can show you error log which i post into irc in the past day 2013-05-13T07:58:55 mips-rtems4.11-gcc -B../../../../../jmr3904/lib/ -specs bsp_specs -qrtems -march=r3900 -Wa,-xgot -G0 -O2 -g -Wall -Wmissing-prototypes -Wimplicit-function-declaration -Wstrict-prototypes -Wnested-externs -march=r3900 -Wa,-xgot -G0 -o capture.exe init.o test1.o 10:02 2013-05-13T07:58:56 weiY (.text+0x8): undefined reference to `__stack_chk_guard' 10:02 2013-05-13T07:58:56 weiY (.text+0xc): undefined reference to `__stack_chk_guard' 10:02 2013-05-13T07:58:56 weiY (.text+0x68): undefined reference to `__stack_chk_guard' 10:02 2013-05-13T07:58:56 weiY (.text+0x88): undefined reference to `__stack_chk_fail' 10:02 2013-05-13T07:58:56 weiY (.text+0x9c): undefined reference to `__stack_chk_guard' 2013-05-13T08:00:55 sebhub, in regards to the newlib reentrancy change over is newlib-cvs-13-May-2013 ok ? 2013-05-13T08:01:19 weiY, you are using the RPM on Ubuntu right? 2013-05-13T08:01:21 no, please use the version from today 2013-05-13T08:02:01 or some kind of deb package produced from somewhere? 2013-05-13T08:02:03 or at least 2013-05-07 2013-05-13T08:02:18 gedare, when i use RPM it report error. so now i have build RSB and it seems ok 2013-05-13T08:02:48 yeah i remember you discussing it here. 2013-05-13T08:03:26 well, good you have a working toolchain now. 2013-05-13T08:03:46 yeah, now the problem is disappared 2013-05-13T08:04:13 but there is also some issue when i build RSB, though this issue is not a really problem 2013-05-13T08:04:15 sebhub, what is today for you ? It is 13 May 2013 for me. :) 2013-05-13T08:04:26 weiY, what issue ? 2013-05-13T08:05:06 oh, confused the month 2013-05-13T08:05:43 sebhub, ok. I have committed and pushed today in the repo 2013-05-13T08:05:52 when i build the RSB encountered some error and if i rebuild the RSB (the source is not clean), the patches will patch again and so it will show error 2013-05-13T08:05:54 I will add the iconv stuff soon 2013-05-13T08:06:06 *** arvind_khadri has quit IRC (Read error: Connection reset by peer) 2013-05-13T08:06:14 weiY, oh on mips ? 2013-05-13T08:06:36 *** arvind_khadri has joined #rtems 2013-05-13T08:06:36 yeah, there is two patch newlib-2.0.0-mips-memcpy-20130107.diff newlib-2.0.0-mips-memcpy-20130219.diff 2013-05-13T08:06:46 if i rebuild it will patch twice 2013-05-13T08:06:57 so i must patch it manually 2013-05-13T08:07:21 I think you may need to update your RSB ? http://www.rtems.org/pipermail/rtems-tooltestresults/2013-May/000698.html 2013-05-13T08:07:26 I built this today 2013-05-13T08:07:38 i build it yestoday 2013-05-13T08:08:07 http://git.rtems.org/chrisj/rtems-source-builder.git/commit/?id=dd8534a39a1692a59e6a5d532b1bcc751c203c2d 2013-05-13T08:08:21 you build it also for mips? 2013-05-13T08:08:25 Took the patches out then 2013-05-13T08:09:40 *** arvind_khadri has quit IRC (Remote host closed the connection) 2013-05-13T08:10:17 and another issue is when i run the sb-check it is ok. but when i build it will show error "python2.7 is not support" although i have install python2.7 2013-05-13T08:10:21 The mailing list post was a clean build 2013-05-13T08:10:42 and finally i found i have not install python-devel 2013-05-13T08:11:30 *** arvind_k has joined #rtems 2013-05-13T08:12:15 ok, i have seen your patch, you have delete the patch script. it should be ok. but if there are some other patches whether it will also encounter this similiar issue? 2013-05-13T08:16:22 the jmr3904 does not support "ll/sc" instruction? 2013-05-13T08:16:52 when i build the atomic test case it shows the error: /tmp/cc2xBZnr.s:496: Error: Opcode not supported on this processor: r3900 (mips1) `ll $2,32($sp)' 2013-05-13T08:24:27 *** arvind_k has quit IRC (Remote host closed the connection) 2013-05-13T08:24:49 *** arvind_khadri has joined #rtems 2013-05-13T08:26:17 weiY, you would have to tell me which patch is breaking the build. The build log should show this. 2013-05-13T08:28:59 it is not the patch breaking the build. it is should be the problem of patch twice if i encounter a error in first build and rebuild 2013-05-13T08:30:08 it means that if the patches is applied and i donot clean the source and build dir. and when i rebuild it will re-apply the patches 2013-05-13T08:31:43 The build normally clean the source before building. The current 4.11 tools is building from git however there should be a reset --hard before the build 2013-05-13T08:35:48 *** WeiWeiY has joined #rtems 2013-05-13T08:36:02 this source dir will patch the two patches newlib-2.0.0-mips-memcpy-20130107.diff newlib-2.0.0-mips-memcpy-20130219.diff twice 2013-05-13T08:37:11 i use the same build log name, so there is only successful log 2013-05-13T08:37:29 *** weiY has quit IRC (Ping timeout: 276 seconds) 2013-05-13T08:37:33 *** arvind_khadri has quit IRC (Remote host closed the connection) 2013-05-13T08:37:41 *** WeiWeiY is now known as WeiY 2013-05-13T08:37:59 *** arvind_khadri has joined #rtems 2013-05-13T08:38:58 WeiY, regarding the jmr, you may have to consult a processor spec sheet. but if the instruction is not decoding, then probably it is not part of the instruction set that is supported. 2013-05-13T08:39:58 yeah, i will read its spec. but about mips SMP atomic debug it is really a problem 2013-05-13T08:40:15 WeiY, in my RSB those patches have been delete and newlib-cvs is being used. I assume you are building 4.11/rtems-mips 2013-05-13T08:40:55 yeah, i build 4.11/rtems-mips. and what should i build? 2013-05-13T08:41:10 That is fine. 2013-05-13T08:41:31 Can you please create a report and sent it to me using the sb-report command ... 2013-05-13T08:41:35 and it seems that it will not clean the source in the newlib-cvs? 2013-05-13T08:41:56 ../source-builder/sb-report --format=text 4.11/rtems-mips 2013-05-13T08:42:26 if you delete the two patches i think this problem will be not shown 2013-05-13T08:42:38 i have git pull your latest repo 2013-05-13T08:42:45 Ah ok. 2013-05-13T08:42:49 Great 2013-05-13T08:43:18 i just want to say if there is also other patches later whether it will show the same error 2013-05-13T08:44:16 *** noselasd_ has joined #rtems 2013-05-13T08:44:21 Get a report and have a look. I do not think there are. 2013-05-13T08:44:40 i think if you clean the newlib-cvs source code when rebuild RSB it will not show 2013-05-13T08:45:17 ok, if i can re-show this issue i will try and report to you 2013-05-13T08:45:26 *** arvind_khadri has quit IRC (Remote host closed the connection) 2013-05-13T08:45:30 Could be the issue. Do you think the patches where applied to the cvs check out ? 2013-05-13T08:46:50 when i encounter this issue i reverse the patches it will ok 2013-05-13T08:47:32 kiwichris_, on ubuntu, RSB needs most of the same packages as raspbian. 2013-05-13T08:47:32 I had to install ncurses-dev in particular 2013-05-13T08:47:32 um, 12.04 ubuntu i think i am on 2013-05-13T08:47:33 it might be good to list all the major package dependencies somewhere easy to find 2013-05-13T08:47:33 and the names of those packages in various distros 2013-05-13T08:47:34 i know it is a little bit of spaghetti :) 2013-05-13T08:47:38 *** noselasd has quit IRC (Ping timeout: 260 seconds) 2013-05-13T08:47:44 *** arvind_khadri has joined #rtems 2013-05-13T08:48:16 The doco should list them. 2013-05-13T08:48:24 and python-devel is also needed. but when run sb-check it will not show problem 2013-05-13T08:49:08 *** sh has joined #rtems 2013-05-13T08:49:32 *** sh is now known as Guest82017 2013-05-13T08:49:40 The sb-check cannot check for all missing depending libraries. 2013-05-13T08:49:47 That would be difficult 2013-05-13T08:50:46 *** rokka has joined #rtems 2013-05-13T08:50:52 *** the9a3eedi has quit IRC (*.net *.split) 2013-05-13T08:50:55 *** gedare has quit IRC (*.net *.split) 2013-05-13T08:51:15 I suppose we could add a file of these for each distro and then check could have some distro code to run a check 2013-05-13T08:51:23 the command "../source-builder/sb-report --format=text 4.11/rtems-mips" will send you directly? 2013-05-13T08:51:57 yeah, really need this file to check all components 2013-05-13T08:51:59 WeiY, do you think your newlib cvs source tree was patched ? 2013-05-13T08:52:10 yeah, i am sure 2013-05-13T08:52:28 becasue i must reverse the patch to a clean status and rebuild 2013-05-13T08:52:36 you can check my cd sources/cvs/ then into the specific newlib dir and run 'cvs -qn update' 2013-05-13T08:54:29 and what next step 2013-05-13T08:54:36 it will not show anything 2013-05-13T08:55:03 *** the9a3eedi has joined #rtems 2013-05-13T08:55:03 *** gedare has joined #rtems 2013-05-13T08:55:39 and then if i run patch command it will show patch -p1 <../../../../../rtems/patches/newlib-2.0.0-mips-memcpy-20130107.diff 2013-05-13T08:55:39 patching file newlib/libc/machine/mips/memcpy.S 2013-05-13T08:55:39 Reversed (or previously applied) patch detected! Assume -R? [n] 2013-05-13T08:55:52 mean it have already appled patch 2013-05-13T08:58:05 *** noselasd has joined #rtems 2013-05-13T09:00:25 The patch is merged into cvs so you get this 2013-05-13T09:00:50 the patch was present in the RSB because it was not in cvs or we used the 2.0.0 tarball 2013-05-13T09:01:34 WeiY, this is the normal cycle or getting code pushed upstream 2013-05-13T09:02:28 but if the patch was present in the RSB then i should ensure it will clean them when rebuild. right? 2013-05-13T09:02:50 if not it will always show error in the re-build process 2013-05-13T09:03:38 *** arvind_khadri has quit IRC (Remote host closed the connection) 2013-05-13T09:04:39 *** arvind_khadri has joined #rtems 2013-05-13T09:05:24 *** noselasd_ has quit IRC (Write error: Broken pipe) 2013-05-13T09:05:25 *** rokka_ has quit IRC (Write error: Connection reset by peer) 2013-05-13T09:05:25 *** Fallenou has quit IRC (Write error: Broken pipe) 2013-05-13T09:05:42 *** sebhub has quit IRC (Remote host closed the connection) 2013-05-13T09:07:55 *** Fallenou_ has joined #rtems 2013-05-13T09:10:17 The source is cleaned away each time. 2013-05-13T09:10:44 In the case of cvs is may not be however if the cvs -qn update is clear then there is not patch 2013-05-13T09:11:14 If you have pulled the latest the patches are gone 2013-05-13T09:12:25 yeah, it will not show again 2013-05-13T09:12:53 correct 2013-05-13T09:16:08 *** freenix has quit IRC (Remote host closed the connection) 2013-05-13T09:16:56 *** arvind_khadri has quit IRC (Remote host closed the connection) 2013-05-13T09:18:20 *** arvind_khadri has joined #rtems 2013-05-13T09:29:07 *** gedare has quit IRC (Ping timeout: 276 seconds) 2013-05-13T09:45:11 *** the9a3eedi-2 has joined #rtems 2013-05-13T09:46:40 *** the9a3eedi has quit IRC (Read error: Connection reset by peer) 2013-05-13T10:19:01 *** WeiY has quit IRC (Ping timeout: 248 seconds) 2013-05-13T10:25:05 *** antgreen has joined #rtems 2013-05-13T10:32:25 *** S_Somani has joined #rtems 2013-05-13T10:33:26 *** arvind_khadri has quit IRC (Remote host closed the connection) 2013-05-13T10:36:17 *** arvind_khadri has joined #rtems 2013-05-13T10:42:42 *** arvind_khadri has quit IRC (Remote host closed the connection) 2013-05-13T10:43:10 *** arvind_khadri has joined #rtems 2013-05-13T10:56:37 *** Guest82017 has quit IRC (Remote host closed the connection) 2013-05-13T11:07:22 *** S_Somani has left #rtems 2013-05-13T11:34:56 *** Fallenou_ has quit IRC (Read error: Connection reset by peer) 2013-05-13T11:35:05 *** Fallenou has joined #rtems 2013-05-13T11:47:14 *** noselasd_ has joined #rtems 2013-05-13T11:48:54 *** noselasd has quit IRC (Write error: Connection reset by peer) 2013-05-13T11:59:52 *** antgreen has quit IRC (Ping timeout: 246 seconds) 2013-05-13T12:18:35 *** arvind_khadri has quit IRC (*.net *.split) 2013-05-13T12:26:20 *** arvind_khadri has joined #rtems 2013-05-13T12:32:30 *** noselasd_ has quit IRC (Read error: Connection reset by peer) 2013-05-13T12:36:35 *** Fallenou_ has joined #rtems 2013-05-13T12:41:35 *** noselasd has joined #rtems 2013-05-13T12:43:16 *** arvind_khadri has quit IRC (*.net *.split) 2013-05-13T12:43:17 *** Fallenou has quit IRC (*.net *.split) 2013-05-13T12:45:02 *** arvind_khadri has joined #rtems 2013-05-13T12:45:36 *** arvind_khadri has quit IRC (Read error: Connection reset by peer) 2013-05-13T12:46:00 *** arvind_khadri has joined #rtems 2013-05-13T12:53:20 *** noselasd has joined #rtems 2013-05-13T12:57:28 *** MegaAlex|away is now known as MegaAlex 2013-05-13T12:58:47 *** noselasd has quit IRC (Read error: Connection reset by peer) 2013-05-13T13:03:58 *** noselasd has joined #rtems 2013-05-13T13:15:23 *** antgreen has joined #rtems 2013-05-13T13:20:42 *** noselasd has quit IRC (Read error: Connection reset by peer) 2013-05-13T13:23:46 *** noselasd has joined #rtems 2013-05-13T13:37:07 *** monstr__ has quit IRC (Ping timeout: 264 seconds) 2013-05-13T14:15:48 *** verm__ has quit IRC (Read error: Connection reset by peer) 2013-05-13T14:16:25 *** verm__ has joined #rtems 2013-05-13T14:16:43 *** arvind_khadri has quit IRC (Ping timeout: 268 seconds) 2013-05-13T14:18:51 *** arvind_khadri has joined #rtems 2013-05-13T14:32:14 *** arvind_khadri has quit IRC (Remote host closed the connection) 2013-05-13T14:32:46 *** arvind_khadri has joined #rtems 2013-05-13T15:01:31 *** arvind_khadri has joined #rtems 2013-05-13T15:19:07 *** antgreen has quit IRC (Ping timeout: 264 seconds) 2013-05-13T15:24:17 *** verm__ has quit IRC (Changing host) 2013-05-13T15:24:17 *** verm__ has joined #rtems 2013-05-13T16:14:14 *** lcpfnvcy has quit IRC (Read error: Connection reset by peer) 2013-05-13T16:14:24 *** lcpfnvcy has joined #rtems 2013-05-13T16:19:15 *** lcpfnvcy has quit IRC (Ping timeout: 256 seconds) 2013-05-13T16:32:01 *** lcpfnvcy has joined #rtems 2013-05-13T17:00:48 *** arvind_khadri has quit IRC (Read error: Connection reset by peer) 2013-05-13T17:01:51 *** arvind_khadri has joined #rtems 2013-05-13T17:41:03 *** MegaAlex is now known as MegaAlex|away 2013-05-13T18:34:39 *** the9a3eedi-2 has quit IRC (Ping timeout: 245 seconds) 2013-05-13T18:58:35 *** the9a3eedi has joined #rtems 2013-05-13T20:04:23 *** freenix has joined #rtems 2013-05-13T20:09:16 hmmm.. in the RTL code.. in libbsd/include/dlfcn.h, line 59. Is that supposed to say __aconst? Because I'm getting a syntax error, and I can't find anything on what __aconst is supposed to be 2013-05-13T20:09:27 it's a typo for __const right? just making sure before I send a patch 2013-05-13T20:12:16 hi, the9a3eedi. you can find the macro in file libbsd/include/sys/cdefs.h line 141 2013-05-13T20:15:48 I see.. I will keep looking 2013-05-13T20:16:50 Does the patch resolve your problem? not just when compiling but loading. 2013-05-13T20:18:15 haven't patched it yet. I don't think it should be necessary since it's supposed to be defined 2013-05-13T20:19:30 the error I'm getting while compiling is rtl\libbsd\include/dlfcn.h:59: error: expected '=', ',', ';', 'asm' or '__attribute__' before 'char' 2013-05-13T20:21:44 I did not encounter this error. your toolchain is rtems 4.10? 2013-05-13T20:22:24 yes 2013-05-13T20:22:39 ok. I will try it now. 2013-05-13T20:23:04 I'm running windows, mingw, trying to build for sparc 2013-05-13T20:41:13 hi, the9a3eedi. I do not have windows environment. I just test it for arm and no error msg. 2013-05-13T20:41:19 on Linux 2013-05-13T20:41:37 I see. I get the feeling that these issues I'm getting is because I'm on Windows 2013-05-13T20:42:08 I tried removing __aconst and the build continued, but then it failed during a "stat" command 2013-05-13T20:42:26 it couldn't handle backslashes properly it seems :/ 2013-05-13T20:43:55 I'll see what I can do 2013-05-13T21:00:10 the9a3eedi, what host are you on ? 2013-05-13T21:00:22 Oh Windows. 2013-05-13T21:00:30 windows 7 64-bit, mingw 2013-05-13T21:00:46 Seems like you have some issues here related to the line endings 2013-05-13T21:00:55 Did you use git-msys ? 2013-05-13T21:01:19 yes 2013-05-13T21:01:27 I haven't had line ending issues before though 2013-05-13T21:02:00 Ah ok. I had some issus with building gcc from git and so forced them all to be lf only. I will dig the setting out, hang on 2013-05-13T21:02:55 I have ?. 2013-05-13T21:02:56 $ cat /c/Program\ Files\ $x86$/Git/etc/gitconfig 2013-05-13T21:02:56 [core] 2013-05-13T21:02:56 symlinks = false 2013-05-13T21:02:56 autocrlf = false 2013-05-13T21:02:56 [color] 2013-05-13T21:03:41 and in my ~/.gitconfig under [core] ... 2013-05-13T21:03:42 eol = lf 2013-05-13T21:03:49 autocrlf = false 2013-05-13T21:06:16 I got autocrlf = true in my Git/etc/gitconfig 2013-05-13T21:08:44 I would change this then set the eof = lf and all files will be lf only. This is important when you get to configure scripts 2013-05-13T21:09:10 Most editors worth using these days can handle this 2013-05-13T21:11:05 yeah. I think I'll do that. I just didnt think line endings would be that big of an issue 2013-05-13T21:37:29 hmm.. I just made sure I got all my eol settings right, though I'm still getting issues. I'm pretty sure EOL is not the problem here. 2013-05-13T21:55:16 does git detect the changing of settings that affect all the files? 2013-05-13T21:55:28 you may need to remove all the files and re-checkout for the setting to do anything 2013-05-13T21:55:41 There are commands to manage this but I just checked out the files again 2013-05-13T21:56:07 i wouldn't trust a single one of them they can be flakey if i can i always blow away the files and re-checkout 100% of the time 2013-05-13T21:56:18 That is what I decided as well. 2013-05-13T22:56:22 verm__, I already did that though, still got issues 2013-05-13T22:56:43 I think this has more to do with the fact that I'm using windows, and not simply an EOF thing 2013-05-13T22:57:03 I have to re-edit some source and waf files in order to get them to compile nicely 2013-05-13T23:02:11 If the files are all lf then it should be ok 2013-05-13T23:02:38 have a look with 'file', ie 'file foo.c' and see what it says 2013-05-13T23:05:48 I tried file *.c, and I got all of them to be ASCII C program text, except for init.c which is ASCII English text 2013-05-13T23:05:58 doesn't say anything about eol thoug 2013-05-13T23:06:00 or open in an editor that will show you the line endings 2013-05-13T23:06:27 well my editor says that the line endings lf in some of the files I've opened 2013-05-13T23:08:23 I solved the __aconst issue by simply removing it. It's not exactly a solution, more like a workaround. My current issue is that when I build, waf tries to use tar with "..\\..\\shell-init", and that causes it to fail 2013-05-13T23:08:59 most probably because tar doesn't like using double backslashes as the directory separator 2013-05-13T23:12:39 Hmm ok. I have not tested this on Windows so there could be problems. 2013-05-13T23:13:50 Just looking ... 2013-05-13T23:14:12 I'll try to fix it if I could, but I'm not very familiar with waf. I'm sure theres some way to specifying path separators explicitly 2013-05-13T23:16:01 I would need to take a closer look 2013-05-13T23:16:15 but I do not have time 2013-05-13T23:16:19 verm__, about ? 2013-05-13T23:20:51 *** arvind_khadri has quit IRC (Ping timeout: 256 seconds) 2013-05-13T23:21:09 I'll try my best. I would really like to get this working 2013-05-13T23:22:48 I just cannot remember the way to get the paths to work 2013-05-13T23:26:21 where is the 'tar' coming from ? 2013-05-13T23:30:49 wscript, line 189 from what I can tell 2013-05-13T23:31:18 it doesn't explicitly specify the path of shell-init, so I assume it's just waf magic doing that 2013-05-13T23:32:13 Yes, I mean on your system 2013-05-13T23:32:17 msys, mingw ? 2013-05-13T23:32:22 git-msys 2013-05-13T23:32:38 Do you have git-msys in your path ? 2013-05-13T23:33:45 I do not have msysgit in my path 2013-05-13T23:33:55 Ok 2013-05-13T23:34:26 If you ever do it _must_ be last 2013-05-13T23:34:39 in a shell so 'type tar' 2013-05-13T23:36:32 *** arvind_khadri has joined #rtems 2013-05-13T23:36:38 I've used "where", because type isn't avaliable in cmd.exe. Where tells me that tar is located in C:\MinGW\msys\1.0\bin\tar.exe 2013-05-13T23:38:14 Oh you are using cmd.exe ? 2013-05-13T23:38:45 I have taken no care is making sure things run from cmd.exe 2013-05-13T23:39:13 hmm... let me try using bash then 2013-05-13T23:39:16 Thanks 2013-05-13T23:40:04 It can be made to work but it takes extra effort 2013-05-13T23:40:33 Does bsdtar exist in the MinGW directory ? 2013-05-13T23:40:58 Also see if MinGW was a tar. 2013-05-13T23:41:04 was -> has 2013-05-13T23:42:06 I can run bsdtar from msys and cmd.exe 2013-05-13T23:42:53 yes, but you need to use a mingw or wndows native version and not an msys version 2013-05-13T23:43:25 I suspect if you start from within sh or bash the msys tar should work cause 2013-05-13T23:43:36 msys will manage the path conversoins 2013-05-13T23:43:58