2011-07-25T00:46:06 *** arvind_k has joined #rtems 2011-07-25T00:46:49 *** arvind_khadri has quit IRC 2011-07-25T01:06:45 *** arvind_k has quit IRC 2011-07-25T01:56:23 *** arvind_khadri has joined #rtems 2011-07-25T01:56:23 *** arvind_khadri has joined #rtems 2011-07-25T04:09:14 *** ghostinthepuppet has joined #rtems 2011-07-25T04:44:08 *** arvind_khadri has quit IRC 2011-07-25T04:46:16 *** arvind_khadri has joined #rtems 2011-07-25T05:28:04 *** ghostinthepuppet has quit IRC 2011-07-25T06:10:35 *** arvind_khadri has quit IRC 2011-07-25T07:10:06 *** kristianpaul has quit IRC 2011-07-25T07:13:37 *** arvind_khadri has joined #rtems 2011-07-25T07:26:40 *** kristianpaul has joined #rtems 2011-07-25T07:51:56 *** dimonov has quit IRC 2011-07-25T07:53:34 *** dimonov has joined #rtems 2011-07-25T08:01:09 *** Shisui has joined #rtems 2011-07-25T08:22:41 *** arvind_khadri has quit IRC 2011-07-25T08:31:59 *** arvind_khadri has joined #rtems 2011-07-25T08:31:59 *** arvind_khadri has joined #rtems 2011-07-25T08:38:53 *** arvind_k has joined #rtems 2011-07-25T08:39:13 *** arvind_khadri has quit IRC 2011-07-25T08:55:06 *** methril_work has joined #rtems 2011-07-25T10:11:11 kiwichris, I just see your message from 14h ago, I apologize ! :-) 2011-07-25T10:27:57 *** ghostinthepuppet has joined #rtems 2011-07-25T11:07:20 *** ghostinthepuppet has quit IRC 2011-07-25T11:10:39 *** ghostinthepuppet has joined #rtems 2011-07-25T11:15:41 *** arvind_k has quit IRC 2011-07-25T11:16:13 *** arvind_k has joined #rtems 2011-07-25T12:50:45 *** ghostinthepuppet has quit IRC 2011-07-25T14:37:14 *** ghostinthepuppet has joined #rtems 2011-07-25T15:03:09 *** ghostinthepuppet has quit IRC 2011-07-25T15:09:08 *** ghostinthepuppet has joined #rtems 2011-07-25T15:10:34 hi all, i have a question about simulators 2011-07-25T15:10:43 i was trying to run the coverage tests 2011-07-25T15:11:02 specifically with ./run_coverage erc32_baseline 2011-07-25T15:11:33 receiving the error "unable to find tsim-erc32 on PATH" 2011-07-25T15:12:49 after checking i didn't find it in the base distribution of rtems, and on the web i saw it's a commercial software that costs quite a bit 2011-07-25T15:13:04 is there some other way to emulate erc32 systems? 2011-07-25T15:13:14 or anyway run the erc32 coverage tests? 2011-07-25T15:48:20 *** gedare has joined #rtems 2011-07-25T15:48:54 *** gedare has quit IRC 2011-07-25T16:05:15 *** ghostinthepuppet has quit IRC 2011-07-25T16:13:19 *** DrJoel has joined #rtems 2011-07-25T16:13:19 *** ChanServ sets mode: +o DrJoel 2011-07-25T16:21:58 *** ghostinthepuppet has joined #rtems 2011-07-25T16:23:11 DrJoel, do you perhaps have a minute? 2011-07-25T16:29:16 sure .. 2011-07-25T16:29:26 but I am scared of ghosts .. 2011-07-25T16:29:36 long over my fear of puppets after the horrible accident LOL 2011-07-25T16:29:49 haha 2011-07-25T16:29:53 thanks for your time 2011-07-25T16:30:08 i'm very much interested in the esa socis project about code coverage analysis 2011-07-25T16:30:25 so i tried to run the existing tests 2011-07-25T16:30:47 specifically with ./run_coverage erc32_baseline 2011-07-25T16:31:14 but it seems i need the tsim package, which has to be bought 2011-07-25T16:31:19 erc32 uses tsim which is a commercial piece of software .. yep 2011-07-25T16:31:38 you want to use either one of the BSPs which runs on skyeye or on qemu 2011-07-25T16:31:59 i have qemu 2011-07-25T16:32:17 qemu stock does not produce trace logs.. If you use qemu from couverture, then pc386 will produce the trace logs 2011-07-25T16:32:18 any suggestion on an arch for playing around a bit? 2011-07-25T16:32:50 pc386 2011-07-25T16:32:58 for coverage .. probably easiest to use pc386 with qemu from couverture http://www.open-do.org/projects/couverture/ 2011-07-25T16:33:00 Shisui, hi 2011-07-25T16:34:38 thanks for the hint. so the socis project on coverage analysis would concern only one initial architecture (e.g. pc386), or do we need gcov/gprof support for all architectures? 2011-07-25T16:36:22 on a sidenote: 'yum install rtems-4.11-i386*' gives me the message "Package rtems-4.11-i386-rtems4.11-gcc-libgcc-4.6.1-5.fc15.noarch.rpm is not signed" 2011-07-25T16:36:45 my understanding is that gcov/gprof output is target independent. So if you can get it to work on one architecture, it should work on all. RTEMS project way is to get one architecture working while watching out for issues which might impact other targets. Then address those on 2nd architecture. But this shouldn't have any (or at least many) 2011-07-25T16:37:13 ghostinthepuppet: post that on the user's list. Ralf apparently made a mistake. 2011-07-25T16:38:05 DrJoel, why do we want gcov/grof output ? 2011-07-25T16:39:04 a few reasons.. gcov in particular has different reports which may provide different insight 2011-07-25T16:39:26 the gcov format is supported by multiple tools (lcov, xcov, etc) so there is that opportunity again. 2011-07-25T16:40:11 Plus those tools may be more capable of generating formal statements of coverage, decision coverage, and MCDC. 2011-07-25T16:40:19 and more accepted when they do it 2011-07-25T16:40:59 Ok. I though they were profiling tools and not coverage. 2011-07-25T16:41:53 gprof is not as directly of interest to me personally but if someone can generate a useful scenario and run it on say qemu (pc386, qemuppc, uc5282 or leon2 at the moment), then they can gather useful profiling data for hand optimizations and compiler profile driven optimizations 2011-07-25T16:42:05 gcov is the standard gcc coverage tool 2011-07-25T16:42:12 gprof is the standard profiling tool. 2011-07-25T16:42:25 Ok I understand. 2011-07-25T16:42:30 Personally more interested in gcov but covoar gathers enough information where it could output gprof also. 2011-07-25T16:42:41 Nice. 2011-07-25T16:42:54 i saw both are already included in the rtems distribution 2011-07-25T16:43:24 so the goal of the socis project wouldn't be to replace the current analysis tools, but just to add gcov/gprof output format to covoar 2011-07-25T16:48:04 phone brb 2011-07-25T16:50:40 hi kiwichris :) 2011-07-25T16:51:10 .. and DrJoel and ghostinthepuppet. 2011-07-25T16:51:46 hi there 2011-07-25T16:52:36 i saw your application file for gsoc, and am working now on my application for esa socis 2011-07-25T16:52:42 esa summer of code 2011-07-25T16:54:10 nice, good luck! 2011-07-25T16:56:31 thanks, hope your work is going well too :) 2011-07-25T16:56:45 btw, your proposed schedule seems very detailed 2011-07-25T16:57:08 did you know all these things before applying, or do they include mentor comments and feedbacks? 2011-07-25T16:59:37 well I took contact with my mentor before submitting my application in order to discover what was really needed for the project, and document myself about the technical part 2011-07-25T17:00:54 the schedule is the result of both feedbacks from various persons on IRC and from the mailing list and from my personal sight of the project 2011-07-25T17:00:55 ok nice, sort of what i'm trying to do here on irc :D 2011-07-25T17:01:10 thanks 2011-07-25T17:01:11 he he :) 2011-07-25T17:01:46 how is your work on the iso9660 going? 2011-07-25T17:01:59 it's always hard to write down a schedule before getting deeper in the project itself :p 2011-07-25T17:02:51 very good, the basics and standard-compliant things of the filesystem are implemented and working, and I'm now dealing with extensions (multi-byte filenames, POSIX attributes, etc.) 2011-07-25T17:07:03 what kind of systems need cdrom file systems? i know rtems for its usage in spacecraft, and they don't seem to have cdrom drives :D 2011-07-25T17:08:18 back.. sorry 2011-07-25T17:08:24 no problem 2011-07-25T17:08:34 ghostinthepuppet: correct.. just another couple of output formats 2011-07-25T17:09:25 right now, it generates a known set of reports in txt and html formats. This would add gcov and gprof files as output 2011-07-25T17:10:10 ghostinthepuppet, it is a standard file system format and good for things like flash 2011-07-25T17:10:26 ghostinthepuppet: I have seen land-based systems which used CD-ROMS for maps.. or to load other types of data 2011-07-25T17:10:47 so not used really operationally as much as to load data before using system 2011-07-25T17:11:11 thanks for the insight! 2011-07-25T17:11:42 lots of systems out there and lots of ways to do things. :-D 2011-07-25T17:11:49 DrJoel: do you remember my project at http://tinyurl.com/3lbsdnm ? 2011-07-25T17:13:48 yes not that you prod me :).. lots to stuff up there.. what are you doing with the system? 2011-07-25T17:15:48 DrJoel: the coverage project also mentions the task of adding more coverage tests 2011-07-25T17:15:51 I just got it working, rtems is running on it and it's awesome. Basically, it's a digital stereo which I can control from a GTK interface 2011-07-25T17:16:02 1. would this be a result from parsing the new gcov output? 2011-07-25T17:16:14 2. how important / how much of the project would it take? 2011-07-25T17:16:52 demute: :) If you would like to have an RTEMS flyer on it, please let me know. You are welcome to add it to the apps on the wiki. In fact, I encourage it. :) 2011-07-25T17:17:21 like, 50% implementing gcov/gprof output, 50% developing new tests (of the total of 2 months, august/september) 2011-07-25T17:17:46 DrJoel: that sounds fun! An RTEMS flyer would be great :-D 2011-07-25T17:17:55 ghostinthepuppet: 1. We know we do not have 100% statement and branch coverage now. The worst possible case is that gcov lets us know we need more tests. So I don't think gcov support changes the need for more tests. 2011-07-25T17:18:38 demute: http://rtems.org/node/70 has the ones done so far. Basically about a page of text (2-4 paragraphs) and a few pictures. 2011-07-25T17:21:04 DrJoel: I will definitely make a flyer. I planned two years ago to run RTEMS on this board and at last, this summer I had time to get it working. 2011-07-25T17:22:30 DrJoel: I have written some i2c code to the at91rm9200 as well, it's maybe a good idea to put it in the repository. 2011-07-25T17:22:32 ghostinthepuppet: 2. gcov support depends primarily on writing a .gcno file (I think) whose format is defined by gcov-io.h in gcc-XXX/gcc subdirectory. I would guess splitting the time between the two is a good guess. But run gcov natively and see what info is in the file 2011-07-25T17:22:46 demute: please submit anything anyone else can use. :-D 2011-07-25T17:23:03 DrJoel: where do I submit it? 2011-07-25T17:23:10 demute: flyers are in scribus.. but if you provide the text and pictures, I am happy to make them. 2011-07-25T17:23:17 demute: just file a Bugzilla PR with it. 2011-07-25T17:25:02 DrJoel: OK thanks for all the information, I think I'm now prepared to write a good proposal. 2011-07-25T17:25:23 well after reading and trying some more things on my machine :) 2011-07-25T17:26:30 :) honestly i haven't used gcov enough to know much about it. But we can figure it out 2011-07-25T17:26:40 DrJoel: ok, will provide you with the text and pictures when I have them ready (a week of vacation first) and I will file a PR also. Talk to you soon and thank you for the great OS you and many others have built. :-D 2011-07-25T17:27:32 demute: it's ok.. plenty of time. .. write the wiki page first. It is easy to grab the text and pics from there 2011-07-25T17:28:31 ok 2011-07-25T17:30:23 bye all! 2011-07-25T17:30:27 bye 2011-07-25T17:30:35 bye 2011-07-25T17:30:35 *** ghostinthepuppet has left #rtems 2011-07-25T17:30:41 time to go to bed, it's 0:30 and I am going up 04:00 to get a boat... bye! 2011-07-25T17:31:36 night.. time for me to head to the gym soon 2011-07-25T17:38:59 odd posix question.. is pthread_join interruptible by signal. What happens if thread is blocked on pthread_join() and a signal is sent which it is willing to receive 2011-07-25T17:41:51 errno == EINT ?? 2011-07-25T17:43:03 specifically says it can't be returned.. gcj project student has a test case that run on linux.. not rtems where it looks like it is interrupted.. runs signal handler.. and returns to blocking.. I think I see hints in opengroup.org that pthread_mutex_lock is supposed to do same. :* 2011-07-25T17:43:06 :( 2011-07-25T17:43:50 Not return EINT when in these calls ? 2011-07-25T17:51:30 not that I can tell from the open group or the student's test case which works under linux. :( 2011-07-25T17:52:16 May need to look at the FreeBSD implementation. That is what I tend to do. 2011-07-25T17:58:33 If a signal is delivered to a thread waiting for a mutex, upon return from the signal handler the thread resumes waiting for the mutex as if it was not interrupted. 2011-07-25T17:58:39 http://pubs.opengroup.org/onlinepubs/007908799/xsh/pthread_mutex_lock.html 2011-07-25T17:58:58 Ok 2011-07-25T18:00:53 my concern is how many of those are in the specs.. want to make a unified sweep of them.. many new test cases. :( 2011-07-25T18:57:48 well there goes the announcement for 4.6.9, 4.7.5, and 4.8.4 legacy releases. :) 2011-07-25T18:58:04 Well done 2011-07-25T18:59:35 just need to get out the money bucket.. it should start falling from the sky any minute.. LOL 2011-07-25T19:03:14 any thoughts on when we should target 4.11 branching? 2011-07-25T19:09:08 i'm finally out of here.. 7pm here... been here since 8am.. time to leave 2011-07-25T19:09:11 *** DrJoel has quit IRC 2011-07-25T20:56:54 *** arvind_k has quit IRC 2011-07-25T22:07:22 *** mwalle has quit IRC 2011-07-25T22:12:33 *** mwalle has joined #rtems 2011-07-25T22:33:39 *** Shisui has quit IRC 2011-07-26T02:03:21 *** arvind_khadri has joined #rtems 2011-07-26T02:04:38 *** arvind_khadri has quit IRC 2011-07-26T04:33:24 *** methril has quit IRC 2011-07-26T04:47:16 *** methril has joined #rtems 2011-07-26T04:50:05 *** mememe has joined #rtems 2011-07-26T04:51:23 *** mememe has joined #rtems 2011-07-26T04:51:40 *** mememe has joined #rtems 2011-07-26T04:53:43 *** mememe has joined #rtems 2011-07-26T04:53:50 *** mememe is now known as lcp 2011-07-26T04:57:17 *** mememe has joined #rtems 2011-07-26T04:58:20 *** mememe has joined #rtems 2011-07-26T04:58:35 *** mememe has quit IRC 2011-07-26T05:01:54 *** lcpfnvcy has joined #rtems 2011-07-26T05:02:36 *** lcpfnvcy has quit IRC 2011-07-26T05:03:27 *** lcpfnvcy has joined #rtems 2011-07-26T05:04:48 *** lcpfnvcy has quit IRC 2011-07-26T05:13:17 *** lcpfnvcy has joined #rtems 2011-07-26T05:40:06 *** lcpfnvcy has quit IRC 2011-07-26T05:54:53 *** dreamer_ has joined #rtems 2011-07-26T07:23:34 hello, everyone :) I am preparing to apply to one of SOCIS projects and I have few questions 2011-07-26T07:23:47 I am going through http://wiki.rtems.org/wiki/index.php/GSoC_Getting_Started at the moment 2011-07-26T07:24:34 and following steps: "Install RTEMS" and build&run hello world example from class-examples are not really required? 2011-07-26T07:25:27 class-examples as separate cvs module is deprecated, yes? it's enough to modify and compile file under rtems/testsuites/samples/hello ? 2011-07-26T07:25:56 (I hope someone will read this today ;) ) 2011-07-26T07:28:45 also, all examples use 4.10, but I installed rtems 4.11 - it shouldn't be problematic, right? 2011-07-26T08:15:12 *** DrJoel has joined #rtems 2011-07-26T08:15:12 *** ChanServ sets mode: +o DrJoel 2011-07-26T08:25:56 dreamer_: ask DrJoel 2011-07-26T08:26:23 * DrJoel hides in a corner.. 2011-07-26T08:38:58 well... where's the question? 2011-07-26T08:42:21 DrJoel: can you check the irc log :) its about 10 lines 2011-07-26T08:42:40 just before you logged in 2011-07-26T08:43:30 http://www.rtems.org/irclogs/html/2011/rtems-30.html 2011-07-26T08:52:43 dreamer_: use examples-v2 cvs module 2011-07-26T08:53:08 dreamer_: are you the one using the Fedora 12 VM? 2011-07-26T08:53:40 DrJoel, question was about installation phase (I figured it out myself, it needs to be done, but it's not required for compilation really); and second one about examples 2011-07-26T08:53:50 no, I use Fedora 15 natively :) 2011-07-26T08:54:03 and class-examples are merged into examples-v2. code may be OK but I think the Makefile's may have a nit that prevents them from working with cvs head 2011-07-26T08:54:04 it's my distro of choice, fortunately :) 2011-07-26T08:54:44 dreamer_: :) I need to post back and find out why that guy downloaded the F12 VM instead of the F14 one. Probably an out of date link somewhere 2011-07-26T08:54:54 but to pass "can you compile example test" I can use hello world from rtems/testsuites/samples/hello as well, right? 2011-07-26T08:55:13 dreamer_: sure.. but focus on getting the application in now.. the deadline is soon. 2011-07-26T08:55:37 yeah, I learned about socis yesterday in the evening, kind-of last minute 2011-07-26T08:56:10 which project are you interested in? Did you see the template we have to help with application? 2011-07-26T08:57:10 gprof or gcof 2011-07-26T08:57:32 but I wonder why no valgrind? it's more popular as profiling tool, I guess 2011-07-26T08:57:58 ok,, ask questions.. and look in irc log for past couple of days.. some discussions on it yesterday I think 2011-07-26T08:59:04 I have no experience with gnu profiling tools, but hey, let's learn something new, right? :) 2011-07-26T08:59:21 I will, thanks 2011-07-26T09:01:52 Check out http://xview.net/papers/gcov/code_coverage_gcov.pdf for a small intro to gcov 2011-07-26T09:02:31 qemuppc or pc386 using couverture variant of qemu are the best for you to do coverage on. sparc BSPs using a commercial simualtor for the data 2011-07-26T11:21:44 *** Shisui has joined #rtems 2011-07-26T12:14:15 DrJoel, after reading code, documentation and playing a bit I decided, that http://wiki.rtems.org/wiki/index.php/UseHashOrMapInNotepadsAndKeys project may be more appropriate for me; let me describe if I understand this project properly: 2011-07-26T12:14:30 there are 2 separate apis for key/value storage retrieval: 2011-07-26T12:14:48 one is called "notepads", it consists of: rtems_task_set_note / rtems_task_get_note calls 2011-07-26T12:15:15 second one is posix api: pthread_{key_create,key_delete,setspecific,getspecific} 2011-07-26T12:15:51 currently these are implemented using C arrays, with size of these arrays being specified during compilation of rtems 2011-07-26T12:16:15 so users can adjust this size for need of their particular application 2011-07-26T12:16:58 and goal of this project would be to offer alternative implementation(s), with better memory complexity and preferably with same time complexity 2011-07-26T12:17:34 using of http://uthash.sourceforge.net/ is suggested, but different implementations may be more suited to certain usecases 2011-07-26T12:18:30 e.g. in certain scenarios hashmap based on splay tree may be efficient (when memory is limited and there is big number of possible values, but there is subset of very frequently accessed keys) 2011-07-26T12:18:46 or in other scanarios http://en.wikipedia.org/wiki/Dynamic_perfect_hashing will work a lot better 2011-07-26T12:19:22 of course it wouldn't hurt to see how different *libc implementations handle this api 2011-07-26T12:21:34 *** antgreen has joined #rtems 2011-07-26T12:24:02 *** peerst has left #rtems 2011-07-26T12:24:26 *** dreamer_ is now known as dreamer__ 2011-07-26T12:28:46 that's the task description.. the idea is that (especially for pthread keys), you want something that is flexible. notepads are not quite as critical since they are completely optional now 2011-07-26T12:35:49 hmm, from sidelines using uthash seems like reasonalble first step to familiarize myself with rtems codebase, style, etc; I see, that there are time benchmarks and other tests for these apis already implemented 2011-07-26T12:36:33 are there other requirements? I am not familiar usage scenarios in real time applications... 2011-07-26T12:36:40 with* 2011-07-26T12:47:53 there is a GSOC project to do posix benchmarks for RTEMS. There is a benchmark for the notepads in one of the tmtests.. best to use powerpc/psim for this type of benchmarking because it reports instructions executed. 2011-07-26T12:48:28 One issue to be aware of is that we want 100% test coverage -- especially for code that is in the "core" of RTEMS. 2011-07-26T12:48:51 dreamer__: yep on first steps. Have you built RTEMS and done the hello world exercise listed for SOC students? 2011-07-26T13:32:43 *** ghostinthepuppet has joined #rtems 2011-07-26T13:38:09 *** ghostinthepuppet has quit IRC 2011-07-26T14:58:25 *** dreamer_ has joined #rtems 2011-07-26T15:42:25 hm.. iconv is driving me crazy 2011-07-26T15:49:17 * DrJoel has never used those methods 2011-07-26T15:49:20 what's happening 2011-07-26T15:52:15 well it's not iconv in itself, it's the function call 2011-07-26T15:52:29 it mess my parameters and I can't explain why 2011-07-26T15:53:25 e.g., I call the function like iconv_open("string1", "string2"), and when I debug with GDB this is not those parameters which are on the stack 2011-07-26T15:53:45 If you put the strings in oversize arrays.. does it help 2011-07-26T15:54:30 ohhhh.. they are const char *.. should be OK 2011-07-26T15:56:34 yep 2011-07-26T15:57:23 I even tested with allocating the string on the heap, same results 2011-07-26T15:58:59 can you step into it? pointing the source at the newlib source untar'ed? 2011-07-26T16:01:53 it's the problem, it's the call which is defunct 2011-07-26T16:02:09 take a look at this short GDB session : http://pastebin.com/1ggw58Zr 2011-07-26T16:02:16 around lines 34-39 2011-07-26T16:02:43 I call the iconv_open(const char*, const char*) with two static strings 2011-07-26T16:03:18 and when I step into the function call, the parameters are messed up 2011-07-26T16:08:33 is that the entire program? 2011-07-26T16:08:46 doh.. send me the program.. 2011-07-26T16:09:09 yep, I reduced it to its minimal form .. 2011-07-26T16:11:31 email it to me.. and I will take a look 2011-07-26T16:17:15 I clean it up a bit 2011-07-26T16:17:59 it will help me and whoever. :-D 2011-07-26T16:24:43 ok I think it's good. I've verified the problem is still present (just in case ... :D ) 2011-07-26T16:24:55 do I send it to you personally or to the gsoc ML ? 2011-07-26T16:25:06 personally is fine 2011-07-26T16:25:10 ok 2011-07-26T16:28:31 sent :p 2011-07-26T16:30:37 the program stops and wait for GDB at startup, since the problem cannot be seen otherwise 2011-07-26T16:55:46 I am running it on sis.. not needed.. 2011-07-26T17:01:13 does the problems appears too ? 2011-07-26T17:01:32 i only tested it on i386 2011-07-26T17:01:54 the problem as best I can tell is that find_alias() called from newlib/libc/iconv/lib/aliasesi.c:210 is not finding any conversions. There may be some initialization or table we are missing. Can you download newlib-1.19.0.tar.gz and poke around in that code. I don't think we have conversion tables. 2011-07-26T17:01:55 what opt option, eg -02 etc ? 2011-07-26T17:01:57 yep 2011-07-26T17:04:54 in fact when I ran my tests the "find alias" step was not reached at all because of the parameters mess 2011-07-26T17:06:01 the conversion from "ascii" to "latin1" may not exist at all, it was two charset supported by iconv for testing purposes 2011-07-26T17:06:40 looking at the newlib source, I see that linux has a fair amount of code for iconv .. replaces everything as best I can tell. May be time to ask on newlib list if the generic code works for anyone, anywhere 2011-07-26T17:07:02 sys/linux/iconv has iconv.c and iconv_open.c. which makes me suspicious 2011-07-26T17:07:30 i need to head home.. kiwichris can you ask about this on newlib? 2011-07-26T17:09:22 Linux iconv support is much complex AFAIK, it handles some conversions more efficiently thanks to direct conversions. Newlib doesn't do that because of the library size, every conversion is made with UCS-4 as an intermediate charset 2011-07-26T17:10:22 there looks to be some initialization missing. Since find_alias is failing. Worth asking them. And looking at the newlib code. 2011-07-26T17:10:27 * DrJoel heads out the door 2011-07-26T17:10:30 *** DrJoel has quit IRC 2011-07-26T17:48:20 *** DrJoel has joined #rtems 2011-07-26T17:48:20 *** DrJoel has joined #rtems 2011-07-26T17:48:20 *** ChanServ sets mode: +o DrJoel 2011-07-26T17:59:00 *** DrJoel has quit IRC 2011-07-26T19:24:35 *** panzon has quit IRC 2011-07-26T20:10:13 *** xiangfu has joined #rtems 2011-07-26T20:24:54 *** panzon has joined #rtems 2011-07-26T20:30:57 *** methril has quit IRC 2011-07-26T20:44:28 *** methril has joined #rtems 2011-07-26T23:19:37 *** Shisui has quit IRC 2011-07-27T02:28:05 *** lcpfnvcy has joined #rtems 2011-07-27T02:35:30 *** panzon has quit IRC 2011-07-27T02:52:35 *** kiwichris has quit IRC 2011-07-27T02:57:38 *** panzon has joined #rtems 2011-07-27T03:33:38 *** ghostinthepuppet has joined #rtems 2011-07-27T03:46:09 *** panzon has quit IRC 2011-07-27T03:52:47 *** ghostinthepuppet has quit IRC 2011-07-27T03:55:32 *** ghostinthepuppet has joined #rtems 2011-07-27T04:13:24 *** ghostinthepuppet has left #rtems 2011-07-27T04:29:12 *** xiangfu has quit IRC 2011-07-27T07:15:10 *** antgreen has quit IRC 2011-07-27T08:06:57 *** methril_work has quit IRC 2011-07-27T08:07:07 *** methril_work has joined #rtems 2011-07-27T08:48:17 *** DrJoel has joined #rtems 2011-07-27T08:48:17 *** ChanServ sets mode: +o DrJoel 2011-07-27T09:09:24 *** panzon has joined #rtems 2011-07-27T09:27:52 *** antgreen has joined #rtems 2011-07-27T10:45:11 *** antgreen has quit IRC 2011-07-27T11:31:14 *** Shisui has joined #rtems 2011-07-27T12:27:28 *** Shisui has quit IRC 2011-07-27T12:30:52 *** Shisui has joined #rtems 2011-07-27T12:43:40 *** Shisui has quit IRC 2011-07-27T13:24:42 *** DrJoel has quit IRC 2011-07-27T15:47:12 *** DrJoel has joined #rtems 2011-07-27T15:47:12 *** ChanServ sets mode: +o DrJoel 2011-07-27T16:57:29 *** DrJoel has quit IRC 2011-07-27T17:40:41 *** panzon has quit IRC 2011-07-27T17:55:45 *** panzon has joined #rtems 2011-07-27T20:01:35 *** kiwichris has joined #rtems 2011-07-27T20:24:43 *** xiangfu has joined #rtems 2011-07-27T21:30:24 *** dreamer_ has quit IRC 2011-07-27T21:36:21 *** L84Supper has joined #rtems 2011-07-27T21:37:10 when id DrJoel usually about? 2011-07-27T21:37:16 id/is 2011-07-27T22:24:40 24 hours 2011-07-27T22:24:44 err wait 2011-07-27T22:24:46 i mean 12 hours 2011-07-27T22:36:17 thanks 2011-07-27T22:36:40 we're getting closer to our ARM project 2011-07-27T22:37:16 looks like ARM Cortex M3 core with several peripherals 2011-07-27T22:38:35 128K flash and 16KB of SRAM to fit everything into 2011-07-27T23:06:12 *** xiangfu has quit IRC 2011-07-27T23:26:36 sounds nice 2011-07-28T00:06:02 the larger part has 256KB flash and 64KB SRAM 2011-07-28T00:33:00 *** monstr-rtems has joined #rtems 2011-07-28T03:18:13 *** xiangfu has joined #rtems 2011-07-28T04:36:57 *** xiangfu has quit IRC 2011-07-28T08:45:50 *** antgreen has joined #rtems 2011-07-28T08:46:13 *** pzagor has joined #rtems 2011-07-28T10:00:28 *** monstr-rtems has quit IRC 2011-07-28T10:18:46 *** dreamer__ has quit IRC 2011-07-28T11:03:10 *** monstr-rtems has joined #rtems 2011-07-28T12:01:48 *** arvind_khadri has joined #rtems 2011-07-28T12:51:18 *** arvind_khadri has quit IRC 2011-07-28T14:28:16 *** monstr-rtems has quit IRC 2011-07-28T16:13:18 *** pzagor has quit IRC 2011-07-28T18:45:50 what's the simplest way to get RTEMS on an ARM cpu? 2011-07-28T18:46:09 I have crosstool-ng working for x86 to arm7 2011-07-28T18:46:27 I didn't see an ARM repo 2011-07-28T18:50:15 or did I http://www.rtems.org/viewvc/rtems/cpukit/score/cpu/arm/ 2011-07-28T18:56:59 http://www.rtems.com/onlinedocs/releases/rtemsdocs-4.10.1/share/rtems/html/started/index.html 2011-07-28T19:05:27 *** antgreen has quit IRC 2011-07-28T19:52:18 *** xiangfu has joined #rtems 2011-07-28T20:30:30 *** xiangfu has quit IRC 2011-07-28T20:33:02 *** xiangfu has joined #rtems 2011-07-28T20:33:24 *** methril has quit IRC 2011-07-28T20:45:13 *** methril has joined #rtems 2011-07-28T20:48:57 *** xiangfu has quit IRC 2011-07-28T20:51:10 *** xiangfu has joined #rtems 2011-07-28T21:51:21 *** xiangfu has quit IRC 2011-07-29T02:30:23 *** xiangfu has joined #rtems 2011-07-29T03:04:12 *** pobara has joined #rtems 2011-07-29T04:51:21 *** xiangfu has quit IRC 2011-07-29T05:35:03 *** pzagor has joined #rtems 2011-07-29T09:49:17 *** antgreen has joined #rtems 2011-07-29T11:38:45 *** DrJoel has joined #rtems 2011-07-29T11:38:45 *** ChanServ sets mode: +o DrJoel 2011-07-29T11:57:06 pzagor: which BSP and qemu version are you using? have you gotten coverage results yourself? 2011-07-29T12:00:25 *** Shisui has joined #rtems 2011-07-29T12:45:45 DrJoel: I'm using quemuppc and Couverture qemu 2011-07-29T12:45:47 QEMU emulator version 0.14.0, Copyright (c) 2003-2008 Fabrice Bellard 2011-07-29T12:48:02 I've just updated another test case on string to number conversion to improve coverage 2011-07-29T12:48:28 I'have to build it once more and if everything is ok I'll submit it 2011-07-29T12:53:17 and we made a minor change to a sim-script, right? 2011-07-29T12:53:58 yes, you have to change "-trace" to "--exec-trace" 2011-07-29T12:54:42 if [ ${coverage} = yes ] ; then 2011-07-29T12:54:42 COVERAGE_ARG="-exec-trace ${1}.cov" 2011-07-29T12:54:42 fi 2011-07-29T12:54:49 in qemuppc.in? 2011-07-29T12:54:55 yes 2011-07-29T12:55:22 a gsoc student thought the trace file format changed but you and I think it was just invocation 2011-07-29T12:55:39 I need to upgrade the versino of qemu used for qemuppc here 2011-07-29T12:56:09 weel, when i change it it works for me :) 2011-07-29T12:57:48 oh, and there should be "--exec-trace" not "-exec-trace" 2011-07-29T13:00:22 thanks.. now to install it on a couple of machines. We are careful to build simulators locally, pick an executable for each BSP and not update until we need to. Argument change is a need to. :( 2011-07-29T13:01:15 sounds like a boring job :( 2011-07-29T13:04:50 not fun but controlling the configuration is important. If an upgrade breaks RTEMS, we stick with the older simulator until we resolve the issue 2011-07-29T13:08:34 You're in US, right ? 2011-07-29T13:10:49 yes. Huntsville Alabama .. home of NASA Marshall .. :-D 2011-07-29T13:12:50 generally speaking we are across the highway from this: http://maps.google.com/maps?q=huntsville+space+and+rocket+center&hl=en&ll=34.71127,-86.653236&spn=0.002734,0.004715&sll=37.0625,-95.677068&sspn=44.658568,77.255859&t=h&z=18 2011-07-29T13:13:47 check out that the lovely display at the space and rocket center. I put you on shuttle pathfinder. zoom out a bit and you can see the Saturn V shadow.. and some other rockets 2011-07-29T13:14:31 nice! 2011-07-29T13:15:05 I've been to Kennedy Space Centre in 2006, they also have Saturn 5 mockup 2011-07-29T13:15:20 and obviously Shuttle mockup to :) 2011-07-29T13:15:37 those things look really amazing 2011-07-29T13:15:59 BTW, which shuttle is that ? 2011-07-29T13:16:30 is it a mockup or real thing? 2011-07-29T13:21:58 it is a fiberglass correctly sized version used to verify fittings on various things like the 747, can it be transported safely from Edwards to Kennedy without going under a bridge too low, etc. 2011-07-29T13:22:25 The standing Saturn V is a replica. The real one is horizontal in a building next to that. 2011-07-29T13:43:12 I guess they are putting real shuttles in museums now :) 2011-07-29T13:44:10 sadly yes.. and there is no replacement US human space transport in place or a program to replace it. They cancelled the program which was a replacement. :( 2011-07-29T13:45:20 well there are some comercial programs, but i guess people at Marshall where not happy about canceling Constelation :( 2011-07-29T13:45:59 that's sad, we won't get far in space without a solid spacecraft 2011-07-29T13:46:03 no.. the people I know here are disgusted.. no technical direction. 2011-07-29T13:48:11 and clear real goal 2011-07-29T13:48:43 ok, I'll have to excuse you and grab something to eat 2011-07-29T13:48:54 it's supper time in Poland :) 2011-07-29T13:49:19 night 2011-07-29T14:58:17 DrJoel: I've just submitted my patch from my today's hunting trip :P 2011-07-29T14:58:40 it covers rest of the code from string to numbers conversion 2011-07-29T14:59:09 I didnt know what are the priorities so I've just taken a random uncovered blocks 2011-07-29T14:59:45 I'm going to sleep now, so you have a great afternoon ! 2011-07-29T15:01:24 *** pzagor has quit IRC 2011-07-29T15:18:48 *** lekernel has joined #rtems 2011-07-29T15:19:30 hi 2011-07-29T15:19:40 are you able to compile zlib in CVS head? I get this: http://pastebin.com/qbSxN9AG 2011-07-29T15:29:25 it builds for me. checking log 2011-07-29T15:29:33 "solved" with a typedef long z_off64_t in zlib.h ... 2011-07-29T15:30:19 hmmm.. I wonder why that is needed.. I built everything last week.. are you using pre-built tools? 2011-07-29T15:30:43 I would ask on the list about that. I built lm32 last week as part of release prep and didn't have that problem 2011-07-29T15:31:33 I'm using the latest lm32 rpms for fedora 15 (which still do not have the divider-enabled multilib btw ....) 2011-07-29T15:34:21 Ping the PR for that.. if you haven't already show a case in newlib code where the code is different and how much it varies. That builds the case for pushing on it harder. 2011-07-29T15:36:40 yeah painful... do I really need to argue why it's better to use the hardware divider instead of software emulation when one is present? 2011-07-29T15:38:22 I can't see why it should NOT be used, except that having many multilibs uses more CPU time during build and more disk space. but they are dirt cheap those days. 2011-07-29T15:38:36 is it used for array indexing? If so, just say that. 2011-07-29T15:39:11 I agree. This is either an oversight or resistance. Just wanting ammo in case it is resistance 2011-07-29T15:56:13 * DrJoel needs to get a new battery for my watch.. see you all soon 2011-07-29T15:56:16 *** DrJoel has quit IRC 2011-07-29T16:18:16 shucks! Just missed DrJoel 2011-07-29T16:37:37 *** methril_work has quit IRC 2011-07-30T00:08:35 *** Shisui has quit IRC 2011-07-30T00:17:23 *** antgreen has quit IRC 2011-07-30T02:49:21 *** xiangfu has joined #rtems 2011-07-30T03:12:48 *** xiangfu has quit IRC 2011-07-30T03:46:48 *** pzagor has joined #rtems 2011-07-30T04:26:10 *** pzagor has quit IRC 2011-07-30T05:01:28 *** dreamer_ has joined #rtems 2011-07-30T07:57:56 *** kristianpaul has quit IRC 2011-07-30T07:59:58 *** kristianpaul has joined #rtems 2011-07-30T07:59:58 *** kristianpaul has joined #rtems 2011-07-30T08:02:42 *** kristianpaul has quit IRC 2011-07-30T08:04:58 *** kristianpaul has joined #rtems 2011-07-30T08:05:56 *** kristianpaul has quit IRC 2011-07-30T08:06:55 *** kristianpaul has joined #rtems 2011-07-30T08:07:38 *** kristianpaul has quit IRC 2011-07-30T08:08:02 *** kristianpaul has joined #rtems 2011-07-30T08:15:46 what's the purpose of set_vector()? why not call rtems_interrupt_catch() directly? 2011-07-30T10:27:34 *** Shisui has joined #rtems 2011-07-30T10:41:07 *** peerst has joined #rtems 2011-07-30T10:49:11 *** Shisui has quit IRC 2011-07-30T10:49:45 *** lekernel has quit IRC 2011-07-30T11:03:14 *** lekernel has joined #rtems 2011-07-30T11:15:32 *** Shisui has joined #rtems 2011-07-30T19:00:53 *** panzon has quit IRC 2011-07-30T19:11:45 *** methril has quit IRC 2011-07-30T19:16:49 *** panzon has joined #rtems 2011-07-30T19:23:31 *** methril has joined #rtems 2011-07-30T22:20:50 *** methril_ has joined #rtems 2011-07-30T22:21:31 *** methril has quit IRC 2011-07-31T00:52:54 *** Shisui has quit IRC 2011-07-31T04:43:42 *** lekernel has quit IRC 2011-07-31T04:47:49 *** lekernel has joined #rtems 2011-07-31T08:02:45 *** antgreen has joined #rtems 2011-07-31T08:57:40 *** DrJoel has joined #rtems 2011-07-31T08:57:40 *** DrJoel has joined #rtems 2011-07-31T08:57:41 *** ChanServ sets mode: +o DrJoel 2011-07-31T09:14:11 *** DrJoel has quit IRC 2011-07-31T10:50:00 *** lekernel has quit IRC 2011-07-31T10:52:28 *** panzon has quit IRC 2011-07-31T11:02:36 *** lekernel has joined #rtems 2011-07-31T11:07:42 *** panzon has joined #rtems 2011-07-31T11:11:37 *** Shisui has joined #rtems 2011-07-31T14:26:37 *** lekernel has quit IRC 2011-07-31T14:27:04 *** lekernel has joined #rtems 2011-07-31T14:30:02 *** lekernel has quit IRC 2011-07-31T14:30:30 *** lekernel has joined #rtems 2011-07-31T15:39:12 *** L84Supper has quit IRC 2011-07-31T16:04:53 *** DrJoel has joined #rtems 2011-07-31T16:04:53 *** DrJoel has joined #rtems 2011-07-31T16:04:53 *** ChanServ sets mode: +o DrJoel 2011-07-31T16:07:21 *** DrJoel has quit IRC 2011-07-31T16:15:19 *** lekernel has quit IRC 2011-07-31T19:14:47 *** peerst has quit IRC 2011-07-31T19:17:09 *** peerst has joined #rtems 2011-07-31T19:50:07 *** panzon has quit IRC 2011-07-31T19:50:41 *** panzon has joined #rtems 2011-07-31T20:58:02 *** antgreen has quit IRC