| 00:06:42 | Russ|werk | hey andersee, you about? |
| 00:06:59 | Russ|werk | I was wondering if ld.so.cache was arch independant? |
| 00:07:11 | Russ|werk | ie, can I use one generated for arm on x86 and vice versa |
| 00:09:07 | andersee | Hmm |
| 00:09:18 | andersee | dunno for sure. Probably not. |
| 00:09:50 | andersee | uClibc doesn't use one by default anymore, so I stopped messing with it. |
| 00:11:16 | fontenot | TomW: i have tinyx working on the tuxscreen - what issues are you seeing? |
| 00:12:28 | andersee | fontenot: So where is make/tinyx.mk? ;-) |
| 00:13:29 | Russ|werk | oh really? |
| 00:13:46 | Russ|werk | so I don't even need one anymore? it will find the libraries on its own? |
| 00:14:27 | fontenot | andersee: it's bzflag's fault. i am addicted to that damn game. |
| 00:14:59 | andersee | Russ: yup. uClibc only uses a cache if you go mucking with the makefile to enable it |
| 00:15:21 | andersee | Russ: ldso/ldso/Makefile that is... |
| 00:15:39 | andersee | Russ: it is mostly useless given the way uClibc works anyways |
| 00:16:00 | Russ|werk | cool, that simplifies things, especially for newbies |
| 00:17:00 | fontenot | andersee: how does uclibc know where to look for libs then ? |
| 00:19:29 | andersee | fontenot: ldso first searches RPATH (if set), then LD_PRELOAD, then the directory in which the shared library loader was found, then LD_LIBRARY_PATH, and finally, it sharts hunting in /usr/lib, /lib, etc. |
| 00:19:50 | andersee | Just about 100% of the time, the needed lib is in the dir where the shared lib loader resides... |
| 00:19:55 | fontenot | aaah, cool. |
| 00:20:19 | fontenot | i didnt know I could use LD_LIBRARY_PATH |
| 00:21:52 | andersee | sure |
| 00:25:32 | andersee | Oops, I lied. LD_LIBRARY_PATH is resolved before the directory in which the shared library loader is found. |
| 00:42:49 | Russ|werk | andersee: the single SYSCTL function seems to be broken on x86 |
| 00:59:52 | scanline | is back (gone 07:13:21) |
| 01:15:37 | jacques | hi ho |
| 01:15:40 | | jg was last seen on #intimate 5 hours, 26 minutes and 43 seconds ago, saying: Yup. [Mon Apr 22 20:48:57 2002] |
| 01:15:40 | jacques | ibot: seen jg |
| 01:15:47 | jacques | faints. |
| 01:36:02 | pmisfit | oh, woe. thermal problems on my only remaining laptop. |
| 01:38:08 | KeyserSoze[BK] | ah. throw it in a bucket of cold water. |
| 01:41:53 | pmisfit | heh. |
| 01:42:34 | pmisfit | contemplates building a custom external heat sink for the thing. |
| 01:42:56 | pmisfit | gotta run, another reboot to screw with the bios power management again. |
| 02:22:28 | scanline | is away: getting blown up |
| 02:47:06 | | jacques was last seen on #intimate 26 minutes and 11 seconds ago, saying: havent looked in a few days tho [Tue Apr 23 03:20:55 2002] |
| 02:47:06 | TomW | ibot seen jacques |
| 02:47:16 | TomW | jacques: Hello |
| 02:48:00 | jacques | hi TomW |
| 02:48:51 | TomW | jacques: I started putting a make/tinyx together.. |
| 02:49:08 | jacques | TomW: you said you had problems? |
| 02:49:46 | TomW | yes, but, let's back up a moment. How did you specify the cross compiler? Edit the xc-011010/Makefile ? |
| 02:50:45 | TomW | ... or did you pass CC=arm-linux-gcc in as a commmand: "make -C build/xc-011010 CC=arm-linux-gcc World"? |
| 02:51:16 | jacques | TomW: no, to build X you need to use config files in config/cf |
| 02:51:30 | jacques | th config files are in the howto |
| 02:51:37 | jacques | then just make World will work |
| 02:52:28 | TomW | what is that skiff/local/ stuff? |
| 02:53:02 | jacques | arg is that in there? |
| 02:53:07 | jacques | let me look |
| 02:53:14 | TomW | I need to edit the /skiff/ stuff to point to where my compiler binaries are? |
| 02:53:25 | jacques | skiff/local is the location of the hh.org toolchain |
| 02:54:06 | TomW | hh.org?? |
| 02:54:14 | jacques | handhelds.org |
| 02:54:19 | TomW | oh, hand-helds, k |
| 02:55:05 | TomW | What if I set it to something like s:/skiff:$ARM-TOOS: and pass $ARM-TOOLS into the make? |
| 02:55:36 | TomW | Then people can edit the main Makefile to tell us where stuff is located to compile the tinyx with? |
| 02:56:03 | jacques | if that's in the config files, they may be olod - I need to check |
| 02:56:05 | jacques | old |
| 02:56:44 | TomW | No problem, I will tackle that from the sources I have. I can make patchfiles to correct for what I am taking off of hh.org site. |
| 02:57:34 | jacques | TomW: where are you seeing skiff/local ?? |
| 02:57:42 | jacques | that's not in my config files linked in the howto |
| 02:57:48 | jacques | step 6 |
| 02:57:49 | TomW | You play with BZFlags damned^H^H^H^H^H^H nice game and I'll work on the make/tinyx. ;-) |
| 02:58:26 | TomW | /home/tom/TuxScreen/buildroot-tux/build/xc-011010/config/cf/cross.def |
| 02:58:32 | TomW | it is in: /home/tom/TuxScreen/buildroot-tux/build/xc-011010/config/cf/cross.def |
| 02:58:41 | TomW | sorry. |
| 02:58:58 | jacques | those are old and for the ipaq |
| 02:59:16 | jacques | use my config files, tho you will still have to change the path to the xcompile toolchain |
| 03:00:08 | TomW | Well, they were the links I took from the Wiki howto. I chose to offload the bz2 image, rather than from the CVS. The wiki told me to replace three files in the config/cf dir and gave me html links to the new files. |
| 03:00:46 | jacques | right, and I just looked at those files - skiff/local is not in them |
| 03:01:41 | TomW | opps, looks like I have problem in my make/tinyx where it is *not* taking your config/cf files off that site. My mistake. |
| 03:01:42 | jacques | but an even worse path is |
| 03:01:58 | jacques | if we can come u pwith the path we want as default, I can change those config files |
| 03:02:06 | TomW | Ok jacques, let me go with what I have so far and see how far along I get, ok? |
| 03:02:26 | jacques | thse days I use /usr/local/uclibc/bin but that's because I build the c++ toolchain... |
| 03:02:37 | jacques | TomW: OK, let me know if there's anything I can help with |
| 03:02:43 | jacques | gotta eat - BIAW |
| 03:03:22 | TomW | Tnx, BTW, I think that the /usr/local/uclibc/bin is probably a good one to use, then if things get screwed up, we can blame andersee! :-D |
| 03:03:33 | TomW | heh, heh |
| 03:04:16 | jacques | we can blame him either way - the other option is the path to the gcc wrapper in buildroot-tux :-) |
| 03:05:10 | TomW | I thought of that, but Erik may have include paths already setup in the uclibc/bin stuff. It would help simplify the build. |
| 03:06:21 | TomW | jacques: time for me to call it a night, tonites' class wore me out. Later.. |
| 03:32:41 | levi_ | Has anyone tried gtkfb on the tuxscreen? |
| 04:16:10 | KeyserSoze | is the serial port on the tuxscreen /dev/ttySA0? |
| 04:16:12 | KeyserSoze | i don't see a /dev/ttyS0. |
| 04:19:37 | TomW | jacques: Ok, so I couldn't sleep while the ideas were still percolating... |
| 04:21:56 | pcrei | Can anyone point me to resources for setting up to use pivot_root? |
| 04:21:59 | jacques | hi TomW I'm just back from dinner |
| 04:23:00 | jacques | pcrei: hmm the ppl on #handhelds.org and #intimate know about that too, in case nobody here can help you |
| 04:23:46 | pcrei | jacques: Thanks :) |
| 04:27:37 | TomW | jacques: I need XFree86-devel installed, correct? |
| 04:34:18 | jacques | TomW: no, everything you need should be in the xc dir |
| 04:35:23 | jacques | there is one or two errors in the xfree build process which are not seen if you have the include files installed in your system already, but that's taken care of in the howto (I think) |
| 04:36:32 | TomW | hmm. It is having fits finding stdio.h right now. |
| 04:37:28 | jacques | are you using buildroot wrapper or c++ toolchain ? |
| 04:38:24 | TomW | .. meaning $(STAGING_DIR)/bin/ ? |
| 04:38:50 | jacques | ok that's the buildroot wrapper then |
| 04:39:03 | jacques | so stdio.h should be in staging_dir/include |
| 04:39:21 | jacques | and the wrapper should look there already |
| 04:39:26 | TomW | cross.def has #define StdIncDir $(STAGING_DIR)/include |
| 04:40:18 | TomW | I am not sure that $(STAGING_DIR) is getting expanded properly.. Will do that next: specify path instead of using env var. |
| 04:40:21 | jacques | yeah that too :-) |
| 04:40:32 | jacques | good test |
| 04:40:34 | TomW | ... you get burned, you learn.. |
| 04:40:39 | jacques | true |
| 04:43:32 | jacques | millenniums ??? |
| 04:43:37 | jacques | that's not a word |
| 04:43:57 | jacques | damn Reuters |
| 04:45:49 | TomW | $(STAGING_DIR) is not being expanded, minor problem. |
| 04:46:00 | TomW | I can pass that in. |
| 04:49:14 | TomW | stdio.h is in $(STAGING_DIR)/include, but stddef.h is in $(STAGING_DIR)/include/linux. shit! |
| 04:49:53 | jacques | what's wrong with that? |
| 04:51:01 | TomW | $(STAGING_DIR)/include/linux is not in the includes path for files like: GetDflt.c (etc) |
| 04:52:44 | jacques | hmm i dont remember having that problem |
| 04:53:02 | KeyserSoze | is having trouble opening a serial port in C. |
| 04:53:23 | KeyserSoze | I tried: fd = open("/dev/ttyS0", 0_RDWR | 0_NOCTTY ); |
| 04:53:34 | jacques | stddef.h should either be included by something in include or be #include <linux/stddef.h> in the source |
| 04:53:47 | jacques | i think it's the former |
| 04:53:56 | KeyserSoze | where fd is an int. the compiler gives a parse error on the "0_" words. |
| 04:54:08 | TomW | problem is in lib/X11/Xlib.h it refers to #include <stddef.h) rather than <linux/stddef.h> |
| 04:54:28 | sjm | KeyerSoze: 'Oh' not 'zero' |
| 04:54:58 | TomW | What about X_WCHAR? do you define that? (see: lib/X11/Xlib.h) |
| 04:55:55 | TomW | jacques: How about this? Tar the tinyx dir that you have, DCC it to me, then I will build a makefile to recreate it? |
| 04:56:16 | jacques | TomW: yes, X_WCHAR should be defined by host.def |
| 04:57:06 | jacques | hmm just a sec |
| 04:57:27 | jacques | X_WCHAR is not defined explicitly, but -DUSE_XWCHAR_STRING is |
| 04:58:47 | jacques | TomW: I cant DCC to you - our firewall blocks it |
| 04:59:08 | TomW | Right, if X_WCHAR is not defined then the lib/X11/Xlib.h will default to a typedef. |
| 05:04:50 | jacques | actually I think it's the other way around |
| 05:08:06 | KeyserSoze | sjm: thanks. i will change that. the odd thing is, it is code I checked the manpage, and it uses a zero. |
| 05:09:06 | KeyserSoze | nevermind. the man page is right. |
| 05:09:06 | KeyserSoze | it was code I got off the web that was wrong. |
| 05:09:06 | KeyserSoze | thanks again. |
| 05:11:40 | jacques | i mtrying to find where X_WCHAR might be defined, but havent found it yet |
| 05:12:47 | jacques | TomW: did you see this: |
| 05:12:50 | jacques | #ifndef X_WCHAR |
| 05:12:50 | jacques | #ifdef X_NOT_STDC_ENV |
| 05:12:50 | jacques | #ifndef SCO324 |
| 05:12:50 | jacques | #ifndef ISC |
| 05:12:50 | jacques | #define X_WCHAR |
| 05:12:57 | jacques | #endif |
| 05:12:57 | jacques | #endif |
| 05:12:57 | jacques | #endif |
| 05:12:59 | jacques | #endif |
| 05:13:11 | jacques | so, I wonder where X_NOT_STDC_ENV would be defined |
| 05:14:05 | TomW | yup. All I can tell you is that the #include file stuff is badly fscked up. Wait, I will show you the search it takes.. |
| 05:15:06 | TomW | ../../config/makedepend/makedepend: warning: FSWrap.c (reading Xlcint.h, line 75): cannot find include file "stdarg.h" |
| 05:15:06 | TomW | not in ../../stdarg.h |
| 05:15:07 | TomW | not in ../../exports/include/stdarg.h |
| 05:15:07 | TomW | not in /usr/X11R6/include/stdarg.h |
| 05:15:07 | TomW | not in ../../include/extensions/stdarg.h |
| 05:15:07 | jacques | TomW: yes I have been ranting about how screwed the X build system is for literally years now |
| 05:15:08 | TomW | not in ../../lib/xtrans/stdarg.h |
| 05:15:10 | TomW | not in /stdarg.h |
| 05:15:12 | TomW | not in /home/tom/TuxScreen/buildroot-tux/build/stage/include/stdarg.h |
| 05:15:14 | TomW | not in /home/tom/TuxScreen/buildroot-tux/build/stage/include/linux/stdarg.h |
| 05:15:16 | TomW | not in /stdarg.h |
| 05:16:36 | TomW | let me try something else: I am compiling from buildroot-tux with 'make -C build/xc011010 World', maybe that is the problem? |
| 05:17:05 | jacques | could be |
| 05:17:19 | jacques | might confuse things |
| 05:17:51 | TomW | maybe I need to (cd $(TINYX_SOURCE) ; make World ; cd $(BUILDROOT) ) |
| 05:18:32 | jacques | it's worth a try |
| 05:19:30 | TomW | nope, same problem. |
| 05:21:16 | TomW | I noticed that you used two different paths on config/cf/cross.def and config/cf/iPAQH3600.cf |
| 05:21:16 | TomW | one went to /home/packages the other went to /thunder/... |
| 05:21:57 | TomW | Same staging targets, but different prefixes on the pathname up to the /staging/ part |
| 05:28:07 | jacques | TomW: hmm |
| 05:29:55 | jacques | handhelds.org is slow letting me logon |
| 05:32:03 | jacques | TomW: I just put the config files I am currently using at that location (on handhelds.org) |
| 05:32:03 | jacques | they are in same place as ones I just replaced |
| 05:32:04 | jacques | these are setup for using the c++ toolchain, but I think they are a bit cleaner |
| 05:35:46 | TomW | I am using arm-linux-gcc version 2.95.3 |
| 05:36:49 | jacques | TomW: you are using the buildroot-tux gcc wrapper which in turn uses that gcc right? |
| 05:39:47 | TomW | yes, with andersee's magic. |
| 05:40:21 | jacques | magic ? something new? |
| 05:42:03 | jacques | TomW: does X build for you manually ? |
| 05:42:58 | TomW | eeek! Build X? Me? 8-O |
| 05:43:52 | jacques | heh |
| 05:45:11 | TomW | heh, I thought that you had this business of building X licked... So, I wouldn't have to go through that pain again! I only successfully built X a few times in the past, what a nightmare, I keep to RPMS as much as possible. |
| 05:46:46 | jacques | TomW: I wa just wondering if the howto worked for you, it works for me and other ppl. but if it wont build manually for you, I doubt it will build automatically :-\ |
| 05:47:54 | TomW | jacques: back in the early '70s, a buddy and I used to smoke a lot of Pot and wrote some really wonderous code. But then when we would get up the next day, we would discover that we could replace the 200 lines of wonderously crafted code with 5..10 lines. |
| 05:48:19 | TomW | That is how I think that they construct the code for X, somebody is still smoking the same weed. |
| 05:48:26 | jacques | LOL |
| 05:48:36 | jacques | i tend to agree, having mucked around in it a bit |
| 05:48:54 | TomW | No, I did try the howto. I couldn't build it from there either. |
| 05:49:01 | TomW | same problems. |
| 05:51:23 | jacques | hmmmm |
| 05:52:12 | jacques | you use the config files, set the path to your buildroot-tux/staging/bin, did a make World, and it did not work? |
| 05:52:52 | KeyserSoze | I've been messing around with my buildroot-tux makefile. nothing fancy, just trying to add a few apps to the picogui directory, and remove ones I don't use to save space. Anyawy, after I made my changes in buildroot-tux/make/picogui.mk, I did a make, and it had this error at the end: which: no mkfs.jffs2 in |
| 05:53:58 | KeyserSoze | ah crap. nevermind. i found mkfs.jffs2. i don't know why it wasn't in my path, but I'll jsut add it's location, and that should fix it. |
| 05:54:19 | KeyserSoze | either I'm sleep deprived, or I've gotten in the X developers stash :) |
| 05:56:02 | TomW | heh |
| 05:57:54 | KeyserSoze | lol. and i ''download ramdisk''ed, but didn't flash it afterwards. flashing will help, I bet. |
| 06:02:29 | TomW | jacques: maybe I am too picky? When it complains about not finding the include files, it doesn't stop running, just says that it is an Error, but keeps running. |
| 06:03:10 | jacques | TomW: dont tell me this is in the makedepend step |
| 06:03:53 | TomW | jacques: err... It is building now. :-/ |
| 06:04:18 | jacques | TomW: ok, let's see how far it gets. I bet I have workarounds for all the remaining issues |
| 06:04:44 | TomW | you're gonna hit me when you find out how I made it work... |
| 06:05:36 | jacques | well i cant do too much damage over irc |
| 06:05:41 | jacques | so you might as well tell me |
| 06:06:25 | TomW | ok. I stopped passing in the value for $(STAGING_DIR), edited the files and changed it to a hardcoded path. |
| 06:07:26 | TomW | I can use a sed script on the files to set that pathway to what the user has (STAGING_DIR) and then it can be hardcoded. |
| 06:07:42 | jacques | yeah that might be necessary |
| 06:08:01 | jacques | who knows what the X build system is doing... |
| 06:08:42 | TomW | colormap.c |
| 06:08:46 | TomW | still going. |
| 06:09:52 | jacques | you will probably have a problem somewhere in lib/X11 |
| 06:10:29 | TomW | make[2]: Leaving directory `/home/tom/TuxScreen/buildroot-tux/build/xc-011010' |
| 06:10:29 | TomW | make[1]: Leaving directory `/home/tom/TuxScreen/buildroot-tux/build/xc-011010' |
| 06:10:29 | TomW | Tue Apr 23 02:12:52 EDT 2002 |
| 06:10:29 | TomW | Full build of Release 6.5 of the X Window System complete. |
| 06:10:47 | jacques | hmmmm |
| 06:11:00 | jacques | but did it really work |
| 06:11:11 | jacques | go into lib and type 'make' |
| 06:11:12 | TomW | haha! I dunno yet. |
| 06:11:22 | jacques | see if it tries to build anything |
| 06:11:33 | jacques | make World does not stop for errors |
| 06:11:42 | jacques | until it absolutely has to |
| 06:11:46 | TomW | it runs but builds nothing. |
| 06:11:50 | jacques | wow |
| 06:12:00 | jacques | no go into programs/Xserver and make |
| 06:12:11 | jacques | BTW, is there a XFBdev in there? |
| 06:12:32 | jacques | Xfbdev |
| 06:13:40 | TomW | collect2: ld returned 1 exit status |
| 06:13:40 | TomW | make: *** [Xfbdev] Error 1 |
| 06:13:57 | TomW | /home/tom/TuxScreen/buildroot-tux/build/stage/lib/libc.so: undefined reference to `main' |
| 06:13:57 | TomW | collect2: ld returned 1 exit status |
| 06:13:57 | TomW | make: *** [Xfbdev] Error 1 |
| 06:14:05 | jacques | COOL |
| 06:14:09 | TomW | eh? |
| 06:14:17 | TomW | did i fuck up? |
| 06:14:23 | jacques | no that's great |
| 06:14:39 | jacques | tht error is from uClibc and is fixed in latest snaps |
| 06:14:51 | jacques | that's the main() linking error |
| 06:15:11 | jacques | you can link it manually as in the howto, but the best solution is to build buildroot with latest uClibc snapshot |
| 06:15:15 | jacques | then it wont happen |
| 06:15:30 | TomW | What do we do now? Pull from CVS for the XFree source instead of the bz2? |
| 06:15:53 | TomW | clean this mess up a bit? |
| 06:16:03 | jacques | no, if you want the ts to work, we still need to use the 10-10-01 source plus the patch on the wiki |
| 06:16:39 | jacques | i have tried to get the later source working but they really screwed up the X ts driver - I have been trying to talk to jg about it for months |
| 06:16:49 | TomW | right. No, I meant cvs :blah:blah: co -D 011010 XFree86 |
| 06:17:04 | jacques | oooh, yeah well maybe |
| 06:17:08 | jacques | what do you think? |
| 06:17:27 | jacques | the bz2 is probably less total bandwidth usage |
| 06:17:33 | TomW | Yes, lets do it against cvs, try to lose as many extra-special tarballs as we can. |
| 06:17:44 | jacques | ok fine with me |
| 06:17:49 | TomW | Up to you, I just package the stuff. |
| 06:18:04 | jacques | cvs is fine and probably what most will prefer |
| 06:18:24 | jacques | so they can remove the -D 011010 and prove to themselves that later source does not work :-D |
| 06:18:47 | jacques | i have one more test for you |
| 06:18:58 | jacques | go into programs/xdpyinfo |
| 06:18:59 | TomW | ... Then some Brave, Curious soul will remove the date designation, slave for days to get it to work against the latest & greatest XFress release... |
| 06:19:02 | jacques | and do a make |
| 06:19:21 | jacques | i need to know if there are any undefined symbols |
| 06:19:27 | TomW | nothing to be done for all. |
| 06:19:58 | jacques | Cool! |
| 06:20:19 | jacques | so this was a manual build or auto ? |
| 06:20:39 | TomW | Let me clean it up, then I'll post the necessary stuff. Then when you have a chance you get the latest buildroot, test what I have, and we go on from there? |
| 06:21:04 | jacques | yes I will have a chance tonight. :-D |
| 06:21:22 | TomW | It will be auto as in: ( cd $(TINYX_SOURCE) ; make World ; cd $(BUILDROOT) ) |
| 06:21:44 | jacques | that's close enough :-) |
| 06:21:59 | TomW | I have got to get a fresh cigar, then back to work. cul. |
| 06:22:04 | jacques | ok |
| 06:22:06 | jacques | enjoy |
| 06:31:14 | TomW | jacques: I need an ultimate target to build against, what is the file(s) that we are attempting to build as the final result? |
| 06:33:04 | jacques | TomW: at a minimum, the X server and some libs |
| 06:33:19 | jacques | let me chck which libs i have installed on my tuxes |
| 06:33:55 | TomW | Yes, I need a filename to satisfy the rule to do the 'make World'. |
| 06:35:04 | jacques | Xfbdev, libX11.so.6.2, libXext.so.6.4 |
| 06:35:32 | TomW | gotcha. |
| 06:35:36 | TomW | tnx |
| 06:36:07 | jacques | those are what I have on one tux - NOTE that the Xserver does not need them, only apps do |
| 06:36:24 | jacques | does not need the libs, it does need Xfbdev of course :-) |
| 06:37:03 | jacques | and different apps will require different libs, but all will require libX11 and most all will require libXext |
| 06:37:53 | jacques | since space is so limited in the onboard flash, I'm not sure how much we should put in the root image by default |
| 06:38:28 | jacques | TomW: after this, maybe you would like to do a rxvt.mk so ppl can have a nice X terminal too :-D |
| 06:39:01 | jacques | i am running rxvt quite well on the tux, tho it requires a bit of mucking with ptys |
| 06:39:21 | jacques | then we can to fltk.mk, flwm.mk ... |
| 06:43:33 | jacques | SDL.mk ... |
| 06:43:50 | jacques | laughs maniacally. |
| 06:44:36 | jacques | glib.mk, gtk+.mk |
| 06:45:16 | jacques | i have notes on all those BTW |
| 06:45:27 | sjm | jacques: Tried doom under tinyx? I had a go with microwin, but gave up after a few hours, built fine, but kept segfaulting :( |
| 06:45:48 | jacques | tho some require the c++ toolchain so I guess that kinda does not make sense in buildroot... |
| 06:46:03 | jacques | sjm: not yet.... hmm |
| 06:46:11 | jacques | sjm: it's on my list tho |
| 06:46:21 | sjm | cool |
| 06:46:49 | jacques | the bigger screen on the tux opens up more game possibilities |
| 06:46:56 | jacques | tho I need to get mine modded for sound |
| 06:50:44 | KeyserSoze | how does your IR keyboard work for gaming? I can hardly stand using mine for typing, it tends to lose keystrokes. |
| 06:52:14 | sjm | KeyserSoze: I guess it depends a lot on position. I have mine directly in front, about 1" spacing, and never lose a keystroke. I have tried from across the room, but never for extended typing sessions.... |
| 06:52:25 | jacques | KeyserSoze: I had that problem - but I put in fresh batteries and it's great now |
| 06:52:34 | jacques | KeyserSoze: the batts that come with the tux are old |
| 06:52:57 | jacques | i can use mine from about 6 feet away now |
| 06:53:05 | KeyserSoze | i'll check the batteries. |
| 06:53:15 | jacques | actually I have to b careful, because I have multiple tuxes in the same room :-) |
| 06:53:23 | jacques | and I can control all at once |
| 06:53:36 | KeyserSoze | LOL. |
| 06:54:03 | jacques | heh BTW I just got that special edition DVD |
| 06:54:09 | jacques | havent watched it yet tho |
| 06:55:00 | jacques | arg, now only over 200 left... |
| 07:24:28 | sjm | home time, night all |
| 07:49:14 | scanline | KeyserSoze: you should ask BZFlag about it :) |
| 07:49:25 | KeyserSoze | I'll ask tomorrow, when people are awake. |
| 07:49:32 | KeyserSoze | BZFlag: you here? |
| 07:50:06 | TomW | jacques: Getting ready to add files to buildroot-tux CVS.. |
| 07:50:17 | jacques | TomW: whoo-hooo :-) |
| 07:51:00 | KeyserSoze | TomW: jacques: do you guys have experience with copying the filesystem acrosss ethernet, instead of using blob and the serial port? |
| 07:51:29 | TomW | "not I" said Tom. |
| 07:51:58 | jacques | KeyserSoze: not that I recall... I did stuff like that with the kernel on ipaq, but not with filesystem on tux |
| 07:52:08 | TomW | That is something that should be in the Wiki. |
| 07:52:18 | TomW | but it probably isn't.. |
| 07:52:36 | KeyserSoze | scanline says he's done it, but it is unreliable. |
| 07:52:54 | scanline | hi LoneTech |
| 07:55:38 | LoneTech | lo |
| 07:56:21 | LoneTech | seems 3c574tx does NOT work in tuxscreen without further modification |
| 07:56:35 | jacques | TomW: I am heading home - I will be on from there shortly |
| 07:56:47 | LoneTech | driver tries to reset twice, but the card doesn't work (thinks it has ethernet address 00:00:00:00:00:00) |
| 07:56:50 | TomW | jacques: Ok, sources are now commited. |
| 07:56:55 | KeyserSoze | 3c574tx? |
| 07:57:13 | jacques | TomW: cool will you be on in a bit for questions or should I try it and get with you tomorrow? |
| 07:57:41 | TomW | jacques: darned if I know how they will build, I'll hang out an play for a bit longer. |
| 07:57:59 | jacques | TomW: great. leaving here now. |
| 07:58:04 | TomW | cul |
| 07:59:16 | LoneTech | KeyserSoze: 3com pcmcia 100mbps nic, before they called them megahertz. |
| 08:00:12 | KeyserSoze | ah. |
| 08:00:45 | KeyserSoze | i have the netgear FA411 that was recommended on the wiki |
| 08:01:22 | KeyserSoze | is the card a 5V, or is there another problem? |
| 08:01:37 | scanline | I have the FA411. It works fine for me |
| 08:02:24 | KeyserSoze | well, I'll see you guys later. gotta get some sleep. |
| 08:02:57 | LoneTech | Keyser[zzz]: it's 5V, but giving it 5V did not help. |
| 08:03:12 | LoneTech | I've ordered an FA411 |
| 08:04:10 | LoneTech | but I think there's a chance the 3com is a slightly smarter nic than an ne2000. |
| 08:05:29 | fUD | Hey |
| 08:05:39 | LoneTech | lo fud |
| 08:15:05 | fontenot | TomW: OK I'm home |
| 08:29:04 | TomW | fontenot: busybox is not expanding properly, still puzzling it out. |
| 08:30:38 | TomW | fontenot: somehow, busybox.mk is getting a value of bzcat for $(UNZIP). Don't see how it is doing this, weird. |
| 08:31:35 | fontenot | TomW: that is strange |
| 08:31:58 | fontenot | TomW: I'm getting 2.88kB/s downloading the uClibc snap :-( |
| 08:32:28 | fontenot | TomW: my build should be to busybox soon - I'll let you know if I have same problem |
| 08:32:36 | TomW | yes, andersee seems to have something else soaking up his processor. |
| 08:33:33 | fontenot | more like his pipe |
| 08:33:51 | fontenot | TomW: OK I am having same problem as you |
| 08:33:57 | fontenot | looking... |
| 08:33:59 | TomW | good! |
| 08:34:06 | fontenot | yes, that is a good thing |
| 08:34:45 | fontenot | jacques% cd make/ |
| 08:34:45 | fontenot | /home/packages/tuxscreen/buildroot-tux/make |
| 08:34:45 | fontenot | jacques% grep UNZIP * |
| 08:34:46 | fontenot | busybox.mk:UNZIP=bzcat |
| 08:34:46 | fontenot | busybox.mk:UNZIP=zcat |
| 08:34:56 | TomW | it is coming from the define in tinyx.mk. |
| 08:34:58 | fontenot | problem in the busybox.mk file |
| 08:35:43 | fontenot | maybe the logic at the beginning of busybox.mk is flawed |
| 08:36:00 | fontenot | see where it tries to decide which source and unzip to use? |
| 08:36:22 | TomW | No, it is from tinyx.mk, if you comment that def out then busybox unpacks ok. |
| 08:36:48 | fontenot | TomW: that's strange because as a test I tried my first build with tinyx undefined |
| 08:37:19 | TomW | so have I. But all the make files are sourced by the toplevel Makefile. |
| 08:37:24 | fontenot | tinyx is commented out in my Makefile |
| 08:37:46 | fontenot | TomW: where are they sourced? |
| 08:37:49 | TomW | So the define would be parsed. We seem to have a problem where the env var is not being redefined. |
| 08:37:59 | fontenot | include make/*.mk |
| 08:38:04 | fontenot | ack that's not good |
| 08:38:17 | TomW | yup. |
| 08:38:41 | TomW | I will change the tinyx.mk to UNPACK instead of UNZIP. |
| 08:38:52 | fontenot | maybe an undef $UNZIP in busybox.mk ? |
| 08:39:44 | TomW | can you do that? Undef an environment var? |
| 08:40:05 | fontenot | you can in the shell |
| 08:40:10 | TomW | Maybe set it to "", then it would go away? |
| 08:40:19 | fontenot | havent found the right way to do it in the makefile yet |
| 08:40:25 | fontenot | that might work |
| 08:41:13 | fontenot | ARG |
| 08:41:22 | fontenot | I think since tinyx comes after busybox... |
| 08:41:27 | fontenot | see what I mean? |
| 08:41:39 | fontenot | tinyx is defining it *after* busybox.mk |
| 08:43:28 | TomW | I will change the def, do what works, too tired to go for the elegant way of things. |
| 08:43:49 | fontenot | as long as that "include make/*.mk" is in there, we are going to have all sorts of trouble |
| 08:44:01 | fontenot | things are not going to work like ppl expect |
| 08:44:43 | fontenot | also, we *have* to get a mirror for andersee's site |
| 08:44:52 | fontenot | now I'm getting < 2kB/s |
| 08:44:56 | fontenot | downloading tinylogin |
| 08:45:44 | fontenot | ok it ran to completion |
| 08:45:53 | fontenot | now to try it with tinyx :-D |
| 08:46:07 | | Howdy Bub |
| 08:46:07 | BZFlag | hello? |
| 08:46:54 | fontenot | hi BZFlag |
| 08:47:18 | BZFlag | all files in make/*.mk are sourced together. |
| 08:47:18 | BZFlag | so defines must be unique. |
| 08:47:24 | BZFlag | so all defines that are specific to that make should be PACKAGE_foo |
| 08:47:55 | fontenot | BZFlag: yeah I'd like to see a convention like that |
| 08:48:24 | fontenot | TomW: I notice it's still downloading the tarball for X - is that how you want it ? |
| 08:48:25 | TomW | BZFlag: thanks for telling us that, heh. |
| 08:48:33 | fontenot | I thought we decided on cvs ? |
| 08:48:52 | TomW | fontenot: for now, I just want to shake out the bugs, then we can go for the CVS. |
| 08:49:17 | fontenot | TomW: ah yes that;s a good plan |
| 08:49:40 | TomW | fontenot: don't mess with success(I hope). |
| 08:50:12 | TomW | Did you use the daily snapshot? or just ran the buildroot-tux/Make as it comes off the CVS? |
| 08:50:17 | fontenot | small steps, minimize number of variables changed at once :-) |
| 08:50:36 | fontenot | TomW: daily snap for uClibc? yes |
| 08:50:40 | TomW | fontenot: yes, years of getting burned by the Big Change(s). |
| 08:50:52 | fontenot | TomW: I want the Xserver link to work :-) |
| 08:51:13 | TomW | Well, we have the problem with linking main.o, so I would need the daily, right? |
| 08:51:22 | fontenot | TomW: correct |
| 08:51:25 | TomW | ok. |
| 08:51:33 | fontenot | TomW: actually it was fixed days ago, but there has not been a release since |
| 08:52:46 | | andersee was last seen on #tuxscreen 8 hours, 27 minutes and 15 seconds ago, saying: Oops, I lied. LD_LIBRARY_PATH is resolved before the directory in which the shared library loader is found. [Tue Apr 23 01:25:31 2002] |
| 08:52:46 | fontenot | ibot: seen andersee |
| 08:53:26 | TomW | pee-ing data off andersee site. |
| 08:53:33 | fontenot | :-\ |
| 08:53:50 | fontenot | wish I had my co-lo already so I could be a mirror |
| 08:54:01 | fontenot | 2kB/s is painful |
| 08:55:58 | TomW | so was 56k for me, about 5 years ago. They will have to pry my 800K modem from my cold-dead-fingers! |
| 08:57:46 | fontenot | TomW: I think I have a better way to handle the h3600_ts.h thing |
| 08:57:59 | fontenot | a simple patch to ts.c in Xserver source |
| 08:58:12 | fontenot | all it really needs is the typedef for one struct |
| 08:59:14 | fontenot | TomW: also, how do you decide what path to set for the gcc wrapper or c++ toolchain? mine is building with the toolchain - I expected it to use the wrapper in staging |
| 09:00:29 | fontenot | TomW: here is the struct that needs to be patched into top of programs/Xserver/hw/kdrive/linux/ts.c |
| 09:00:33 | fontenot | typedef struct { |
| 09:00:33 | fontenot | unsigned short pressure; |
| 09:00:33 | fontenot | unsigned short x; |
| 09:00:33 | fontenot | unsigned short y; |
| 09:00:34 | fontenot | unsigned short pad; |
| 09:00:34 | fontenot | struct timeval stamp; |
| 09:00:36 | fontenot | } TS_EVENT; |
| 09:00:50 | fontenot | with that, we con't need the h3600_ts.h file at all |
| 09:02:00 | TomW | drat! I moves some stuff and lost my connection. Have to get uClibc all over again. |
| 09:02:25 | fontenot | TomW: maybe I can DCC it to you - I am at home now.... hmm I have a fw here too |
| 09:03:01 | TomW | naw, it'll get there. I want to look into the ts.c stuff. |
| 09:03:14 | fontenot | TomW: ok |
| 09:03:44 | TomW | As to the gcc, I am not doing anything special, just patching your cross.def file. |
| 09:04:50 | fontenot | TomW: hmm, mine is building with /usr/local/uclibc/bin/arm-linux-gcc |
| 09:05:00 | fontenot | which is my c++ toolchain |
| 09:06:07 | TomW | look in the build/staging/usr/bin andersee has a fake gcc in there. I don't know how he does it, but he is referencing the real arm-linux-gcc with some kind of a wrapper. |
| 09:06:52 | TomW | mine build with ../buildroot-tux/build/stage/bin/arm-uclibc-gcc |
| 09:06:56 | fontenot | TomW: yeah that's the buildroot gcc wrapper I was talking about earlier |
| 09:07:30 | fontenot | but the thing is, the arm-linux.gcc in my path is /usr/local/arm/bin, yet somehow buildroot is using /usr/local/uclibc/bin |
| 09:07:35 | BZFlag | TomW: I changed it to TINYX_CAT instead. |
| 09:07:41 | fontenot | almost as if the config files were not being modified at all... |
| 09:07:49 | TomW | BZFlag: ok. |
| 09:08:18 | TomW | fontenot: ?? you mean the cross.def is not being changed? |
| 09:08:30 | fontenot | TomW: did you modify your patch for the latest config files I uploaded ? |
| 09:08:36 | fontenot | maybe that;s the problem |
| 09:09:31 | fontenot | TomW: heh, it still built :-D |
| 09:09:43 | TomW | hoo boy. |
| 09:09:50 | TomW | But, will it run? |
| 09:10:09 | fontenot | TomW: I bet it will, tho if you;re not patching ts.c, the ts wont work |
| 09:10:36 | fontenot | TomW: would you like me to make the config files on intimate.handhelds.org more "patch-friendly" somehow ? |
| 09:11:08 | fontenot | TomW: want me to send the X server to one of my tuxes and try it ? |
| 09:11:20 | TomW | fontenot: one thing the ts.c stuff will enable me to do: get rid of the "locate -n 1 h3600_ts.h" hack in make/tinyx.mk |
| 09:11:33 | TomW | YES! |
| 09:11:39 | fontenot | TomW: yeah, I said that above :-D |
| 09:11:57 | fontenot | ok give me a few |
| 09:12:19 | fontenot | BTW, here is the strip command I use on the Xserver and libs: |
| 09:12:40 | TomW | yes, we will need that as well. tnx. |
| 09:12:45 | fontenot | staging_dir/usr/bin/arm-uclibc-strip -x --strip-unneeded --remove-section=.note --remove-section=.comment Xfbdev |
| 09:13:02 | fontenot | i got that from andersee and it seems to work fine |
| 09:13:13 | fontenot | (I think it;s in the howto anyway) |
| 09:14:25 | TomW | umm, I didn't finish reading the howto, was too caught up in all the details. |
| 09:16:19 | TomW | finally, building tinyx now.. |
| 09:16:43 | fontenot | so you missed the part at the end that says "or instead of all that stuff above, just do this one line build" ? |
| 09:16:50 | fontenot | :-D |
| 09:16:58 | TomW | :-/ |
| 09:18:53 | fontenot | heh, once the onboard flash starts getting full, it takes a long time to send stuff because it has to erase to make space |
| 09:19:58 | TomW | yeah, I usually 'erase ramdisk' before an upload. Ever notice that if you burn the kernel, then burn the ramdisk that df -h reports a lot less freespace in the JFFS? |
| 09:21:21 | fontenot | TomW: I'll have to try that, but I;m not sending the whole root image, just Xfbdev using zmodem |
| 09:21:56 | fontenot | hmm tux is locked since it ran out of flash ... I thought I had enough |
| 09:25:37 | fontenot | rebooted it, sending again |
| 09:25:46 | fontenot | it had been up for > month anyway |
| 09:26:52 | TomW | Oooh! I have a programs/Xserver/Xfbdev.. |
| 09:27:09 | fontenot | :-) |
| 09:27:30 | TomW | so, do I just stick it somewhere in flash and then ./Xfbdev? |
| 09:30:52 | fontenot | TomW: yes, put Xfbdev someplace like /usr/local/bin or whatever |
| 09:30:58 | fontenot | TomW: then Xfbdev -ac & |
| 09:31:12 | fontenot | the -ac disables access control so you can display X apps to your tux remotely |
| 09:31:28 | fontenot | BTW, the server seems to be working fine with the exception of the ts |
| 09:32:17 | fontenot | I am displaying rxvt and flwm to it remotely from my other tux |
| 09:32:40 | LoneTech | using x2x? |
| 09:35:16 | fontenot | LoneTech: no, just -display |
| 09:35:55 | LoneTech | k. was thinking x2x would let you control the pointer. |
| 09:36:20 | fontenot | LoneTech: hwh cool |
| 09:36:23 | fontenot | heh |
| 09:36:35 | fontenot | but I want to control the pointer with the ts :-) |
| 09:36:50 | LoneTech | nods |
| 09:36:54 | LoneTech | worked for me in picogui. |
| 09:37:54 | fontenot | LoneTech: ts is working in picogui now ? excellent |
| 09:39:14 | LoneTech | well, last test I made it was uncalibrated, but that's since scanline added support for _doing_ the calibration |
| 09:40:45 | TomW | fontenot: Hey! did I just "hear" you say the the Xfbdev is working fine? |
| 09:43:33 | TomW | drat! It is getting near daylight outside! |
| 09:45:05 | fontenot | TomW: yes you did |
| 09:45:13 | fontenot | TomW: exceopt the ts of course |
| 09:45:14 | TomW | Whahooo! |
| 09:45:44 | fontenot | TomW: BTW, if you try it and get the "cant open ts device" error, just symlink /dev/ts to /dev/ucb1x00-ts |
| 09:45:46 | TomW | I am still mucking about in the makefile stuff. |
| 09:45:51 | fontenot | it wont work, but at least the server will start |
| 09:46:58 | TomW | finally, downloading stuff now into the Tux. |
| 09:49:42 | fontenot | heh, I'm running Xscreensaver remotely to tux |
| 09:49:51 | fontenot | it's doing "matrix" right now |
| 09:52:01 | fontenot | TomW: did you remember to strip the binary? you will save > 100kB |
| 09:52:38 | TomW | HEYYYYYYYYYY!! I HAVE X RUNNING ON MY TUX! Hee Hee! |
| 09:53:05 | BZFlag | locate: invalid option -- n |
| 09:53:11 | fontenot | congratulations |
| 09:53:17 | fontenot | BZFlag: yeah we are going to get rid of that |
| 09:53:28 | BZFlag | heh |
| 09:53:39 | fontenot | BZFlag: a simple patch to ts.c will make that whole .h file unnecessary |
| 09:54:05 | fontenot | TomW: imagine how nice it will be with the ts working :-D |
| 09:54:20 | fontenot | BTW, I have a better patch than what's on the wiki |
| 09:54:32 | fontenot | the "calibration" values in the wiki one are off |
| 09:54:46 | fontenot | mine arent perfect but are better |
| 09:54:51 | BZFlag | dirs should not be targets, they keep changing dates. |
| 09:55:02 | TomW | fontenot: Ok. I am going to clean up a few things, commit what I have, then turn the touchscreen problem over to you to look after. I have had it for the night and won't be able to get back to this until maybe Wednesday. |
| 09:56:02 | TomW | fontenot: thanks for your work on the tinyx and your assistance with getting the makefile working. |
| 09:56:59 | fontenot | TomW: ok that sounds great - good work! |
| 09:57:34 | fontenot | now that you;ve done the hard stuff with the makefile, I can get the rest working I think |
| 10:00:27 | BZFlag | loves the teamwork. ;-) |
| 10:00:53 | TomW | fontenot: "the hard stuff"? Heh, Makefiles are cake to do. |
| 10:01:36 | fontenot | not if you havent done them in 7 years and were never particularly good at them :-) |
| 10:01:58 | TomW | fontenot: mucking around inside of X to make it run on an embedded box, that's magic! |
| 10:02:22 | fontenot | heh, well the hard stuff tehre was done by jim gettys and keith packard for the ipaq |
| 10:02:40 | fontenot | i just banged my head against it until it worked |
| 10:02:42 | fontenot | :-D |
| 10:02:48 | TomW | heh. :-D |
| 10:03:00 | fontenot | of course, I got it to build vs uClibc :-D |
| 10:03:19 | fontenot | so I can feel good about that |
| 10:05:51 | LoneTech | (: |
| 10:09:19 | TomW | fontenot: commited, it is now all yours, g'nite all. |
| 10:10:00 | BZFlag | for makes, one should do: make;make tinyx;make |
| 10:10:12 | BZFlag | no need to edit the root Makefile. |
| 10:10:57 | fontenot | interesting |
| 10:10:58 | file | appears |
| 10:11:07 | file | I run TinyX on my laptop, it's nice |
| 10:12:09 | fontenot | file: which laptop ? |
| 10:12:15 | file | turtle |
| 10:12:21 | file | it's 75MHz 8MB RAM |
| 10:12:26 | file | 250mb hardrive |
| 10:12:30 | file | color screen though :) |
| 10:12:33 | fontenot | ah, I want to try it on my picturebook |
| 10:12:55 | file | what's the difference between TinyX and real X? |
| 10:13:33 | fontenot | tinyx is missing some feature and thus is smaller and requires less resources |
| 10:13:41 | file | uh huh |
| 10:13:43 | fontenot | but it's only the server - the libs are the same |
| 10:13:58 | fontenot | no dynamically loaded drivers for one |
| 10:14:13 | file | uh huh |
| 10:24:40 | file | pokes people |
| 12:22:41 | LoneTech | got my FA411 |
| 12:41:32 | fontenot | cool that's what I have |
| 12:41:37 | fontenot | they work great |
| 12:42:41 | LoneTech | hmm.. someone got 3c589 working? |
| 12:42:52 | LoneTech | I wonder if something similar can be done for my 574 |
| 12:44:51 | LoneTech | I will try |
| 12:45:00 | LoneTech | if so, I don't need the FA411 |
| 12:49:12 | fontenot | i have a 572BT |
| 12:49:30 | fontenot | think it might need 5V tho |
| 12:51:02 | LoneTech | the 574 should. |
| 12:51:17 | LoneTech | is 572 10mbit? |
| 12:51:29 | fontenot | yeah |
| 12:51:41 | fontenot | hmm, says 10/100 on it |
| 12:51:50 | LoneTech | k |
| 14:30:10 | LoneTech | does the FA411 work on 3.3V? |
| 15:17:12 | prpplague | Morn: morn |
| 15:17:20 | Morn | hi how are you? |
| 15:17:55 | prpplague | Morn: i send you this file, for advice? lol |
| 15:18:06 | prpplague | Morn: i'm fine you? |
| 15:18:14 | prpplague | Morn: break any phones lately? |
| 15:18:39 | Morn | totally! |
| 15:18:42 | Morn | giggles |
| 16:54:50 | pmisfit | heh. from tinyx patch: #include "FSCKINGXSTUFF/include/stdlib.h" |
| 16:54:56 | pmisfit | that can't be right. |
| 16:55:36 | pmisfit | figures what the hell, and types "make" anyway |
| 17:58:40 | pmisfit | hmm. they did some good work on tinyx last nite. |
| 17:59:42 | pmisfit | ok, well, it insists on trying to use /usr/local/uclibc/bin/arm-linux-gcc, which does not exist. |
| 18:00:36 | pmisfit | oops... screaming kid wanting to be fed. no time to play with this today. |
| 18:01:24 | pmisfit | goes off to feed the girlie |
| 18:01:34 | pmisfit | is away: whoops... pmisfit interruptus |
| 18:36:38 | LoneTech | FA411 worked nicely. (: |
| 18:41:59 | LoneTech | now to see if tvgestel's changes work on 3c574 too |
| 18:48:17 | LoneTech | I don't get a MAC address. |
| 18:54:49 | LoneTech | strange |
| 18:55:00 | LoneTech | can pcnet_cs not be built into the kernel |
| 18:55:01 | LoneTech | ? |
| 19:29:19 | | andersee was last seen on #tuxscreen 19 hours, 3 minutes and 48 seconds ago, saying: Oops, I lied. LD_LIBRARY_PATH is resolved before the directory in which the shared library loader is found. [Tue Apr 23 01:25:31 2002] |
| 19:29:19 | NOYB | ibot: seen andersee |
| 19:29:29 | NOYB | yup |
| 19:50:22 | KeyserSoze | anyone use pthreads on their tux? |
| 19:51:11 | KeyserSoze | I get these linker errors:serialdata.o(.text+0x244): undefined reference to `pthread_create' |
| 19:51:12 | KeyserSoze | serialdata.o(.text+0x518): undefined reference to `pthread_join' |
| 19:53:59 | KeyserSoze | mayeb i need to put something in the makefile, so the linker can find pthreads? |
| 19:59:46 | sjhill | did you enable threads in your Config ? |
| 20:01:16 | KeyserSoze | i think i fixed it. i added -lpthread in my Makefile.am |
| 20:30:24 | sjhill | Calling application main() |
| 20:30:25 | sjhill | ---RESOLVING--- |
| 20:30:25 | sjhill | LOAD ADDR: 0x0 |
| 20:30:25 | sjhill | SYM INDEX: 14 |
| 20:30:25 | sjhill | GOTSYM: 13 |
| 20:30:25 | sjhill | LOCAL_GOTNO: 7 |
| 20:30:28 | sjhill | SYM: __uClibc_main |
| 20:30:30 | sjhill | VALUE: 0x2aaf8380 |
| 20:30:32 | sjhill | ---RESOLVING--- |
| 20:30:34 | sjhill | LOAD ADDR: 0x2aaee000 |
| 20:30:36 | sjhill | SYM INDEX: 685 |
| 20:30:38 | sjhill | GOTSYM: 539 |
| 20:30:40 | sjhill | LOCAL_GOTNO: 10 |
| 20:30:42 | sjhill | SYM: main |
| 20:30:44 | sjhill | VALUE: 0x400470 |
| 20:30:45 | sjhill | ---RESOLVING--- |
| 20:30:47 | sjhill | LOAD ADDR: 0x0 |
| 20:30:50 | sjhill | SYM INDEX: 15 |
| 20:30:52 | sjhill | GOTSYM: 13 |
| 20:30:54 | sjhill | LOCAL_GOTNO: 7 |
| 20:30:56 | sjhill | SYM: printf |
| 20:30:58 | sjhill | VALUE: 0x2ab0d370 |
| 20:31:00 | sjhill | Segmentation fault (core dumped) |
| 20:31:04 | sjhill | i am SOOOO close |
| 21:17:47 | levi_ | andersee: Hey, I just got buildroot working, but I ran into a little snag in the picogui stuff. |
| 21:18:35 | andersee | hi levi_ |
| 21:19:01 | levi_ | For some reason, while trying to install tpcal, it tried to run themec and failed. |
| 21:19:22 | levi_ | I got around it by cding to the picogui dir and doing a make install from there. |
| 21:20:18 | levi_ | After that, the root-level make finished fine. |
| 21:21:07 | andersee | Sounds like a buglet in the picogui makefile for sure |
| 21:23:14 | levi_ | What's different in how it's run from the buildroot-tux makefile as opposed to running it directly? |
| 21:27:05 | andersee | Hey Tim |
| 21:27:11 | andersee | How goes the battle? |
| 21:30:00 | prpplague | anyone here got contacts at snapgear? |
| 21:31:07 | andersee | prpplague: I know some folks there |
| 21:33:17 | prpplague | andersee: we've just started using the snapgear stuff, we'd like to get some code changes, but we can't seem to get anyone to return a phone call about it |
| 21:35:24 | andersee | prpplague: Most of the folk that actually do stuff are in australia |
| 21:35:37 | andersee | prpplague: I can provide you with email addresses... |
| 21:37:44 | BZFlag | andersee: slowly. |
| 21:39:16 | prpplague | andersee: thanks that would be great |
| 21:40:08 | cbiffle | If anyone's interested, I've got Opie up on a couple of TuxScreens here. (Opie is a fork of Qtopia with cooler apps. :-) |
| 21:40:50 | andersee | cbiffle: running on top of X? |
| 21:40:53 | file | Opie is l33t |
| 21:44:59 | cbiffle | andersee: Naw, X baaad. Running on framebuffer over QtE. |
| 21:45:10 | cbiffle | andersee: Well, more specifically, framebuffer takes up less flash. :-) |
| 21:45:18 | BZFlag | compaq has qtopia mostly running over X. |
| 21:46:37 | BZFlag | they would like to see font support and antialiasing etc get turned off in Qt/E and use the native X support for those and any other possible features. |
| 21:47:38 | BZFlag | this should get close to the same flash footprint that Qtopia Qt/E takes but include the remote X capabilities. ie: run X apps remote and display on your handheld, and run handheld apps "remote" on your desktop. |
| 21:47:40 | cbiffle | The iPaq users I know are running it against framebuffer directly. |
| 21:47:55 | cbiffle | *shrug* Qtopia's dinky in my image, for the most part. |
| 21:48:15 | BZFlag | yes. as I said, compaq has it "mostly" working. it's not ready for maintream quite yet, but should be soon. |
| 21:52:00 | andersee | is working on the new config system for uClibc |
| 21:52:15 | andersee | Most of libc now compiles |
| 21:59:03 | cbiffle | BZFlag: ahh, groovy! |
| 21:59:27 | cbiffle | andersee: One of my beefs w/ Qtopia so far is that QtE has difficulties with uClibc. i.e. it don't workin'. :-) |
| 22:00:44 | andersee | details? inquiring minds want to know... |
| 22:02:08 | cbiffle | andersee: Well...*shrug* It simply doesn't work. :-) Currently, I suspect it's looking for a version of libc greater than what uClibc is making available. I don't know if there are really missing functions in uClibc or if it just thinks there will be. :-) |
| 22:02:18 | cbiffle | I note that my copy of uClibc is symlinked as libc.so.0. |
| 22:02:37 | andersee | As designed... |
| 22:03:42 | cbiffle | *nod* QtE looks for a higher version, perhaps 6? I'm not at my tuxscreen and telnetd craps out. :-) |
| 22:03:43 | cbiffle | I haven't tried symlinking uClibc into libc.so.6 and running it. :-) |
| 22:03:59 | cbiffle | Incidentally, anyone in channel know what "telnetd: all network ports in use" means out of telnetd? It doesn't mean the obvious -- not all ports are in use. :-) |
| 22:04:13 | cbiffle | I suspect it's a bug with feeling up socket binding availability. |
| 22:05:07 | cbiffle | What are the principal differences between libc and uClibc? |
| 22:07:12 | andersee | cbiffle: http://www.uclibc.org/FAQ.html |
| 22:08:01 | cbiffle | Yaaay faqs. |
| 22:21:46 | cbiffle | So I suppose the main issue with Qt and uClibc seems to be that uClibc used to have trouble with C++. Yes? Does it still? |
| 22:26:45 | andersee | nope. C++ works fine. |
| 22:28:21 | prpplague | andersee: ok cobol link against uclibc next... you up to it?lol |
| 22:28:56 | cbiffle | Yaaay cobol. |
| 22:29:06 | vosque | AND THEN FORTRAN |
| 22:29:12 | vosque | :P |
| 22:29:20 | cbiffle | Naw. It's all about FORTH. You young'uns. :-) |
| 22:29:56 | cbiffle | andersee: I'll try to browbeat QtE into binding against uClibc to get meaningful debug information out. As vosque has pointed out from the next cube over, a QtE/uClibc/busybox/opie system would be damn small. |
| 22:31:04 | cbiffle | It might very well fit into the Tuxscreen's flash. (Though I'm probably wrong.) |
| 22:34:22 | andersee | cbiffle: I have c++ support in the tuxscreen buildroot... Just enable stlport... |
| 22:37:49 | cbiffle | andersee: Ooh, groovy. Will do. It's mostly QtE that I expect to have to beat. |
| 22:38:42 | andersee | cbiffle: see the notes in make/stlport.mk regarding needed CFLAGS... |
| 22:40:19 | cbiffle | andersee: Groovy! Thanks. It's always nice when the maintainers of the software I use every day are real people. :-) |
| 22:40:31 | mallum | BZFlag: http://handhelds.org/~mallum/screenshots/qpe-matchbox-0.2pre3.png is a shot of qpe under X |
| 22:40:54 | scanline | is back (gone 20:18:27) |
| 22:41:57 | mallum | BZFlag: I havn't toyed with it much yet though ... |
| 22:56:55 | cpenguin_meeting | andersee: you're a real people? |
| 22:57:22 | cpenguin_meeting | :) |
| 22:58:19 | sjm | has entered the office |
| 23:04:41 | cpenguin_meeting | loves IRCing during a particularly boring con call |
| 23:14:37 | | fontenot was last seen on #tuxscreen 10 hours, 22 minutes and 56 seconds ago, saying: hmm, says 10/100 on it [Tue Apr 23 13:51:41 2002] |
| 23:14:37 | TomW | ibot, seen fontenot |
| 23:34:24 | KeyserSoze | ack. when i have no program running that uses the serial port, and I use my GPS to send a bunch of data to the port, my tuxscreen reboots. |
| 23:39:19 | cbiffle | Ooooh! Neat! |
| 23:39:53 | cbiffle | Sounds like something overran its buffer to me. Is the GPS using hardware handshaking? |
| 23:40:20 | KeyserSoze | no handshaking. |
| 23:40:31 | cbiffle | That may be the problem. |
| 23:40:35 | cbiffle | What speed are you transmitting? |
| 23:40:38 | KeyserSoze | does that make sense, even when no one has the serial port open? |
| 23:40:44 | KeyserSoze | only 4800, 8N1 |
| 23:40:48 | cbiffle | Wow. Okay. |
| 23:40:57 | KeyserSoze | and it's only doing a few hundred characters a second. |
| 23:41:07 | vosque | you don't have a console still on the serial port, do you? |
| 23:41:22 | KeyserSoze | no, I shut it off. that was my first though, too. |
| 23:41:24 | cbiffle | My suspicion is that, since the kernel maintains a serial receive buffer, the system fills it up and destroys a portion of kernelspace. However, at those speeds, I doubt that's right. |
| 23:41:32 | cbiffle | shuts up now. |
| 23:41:45 | KeyserSoze | wait. did you say console, or terminal? |
| 23:41:55 | KeyserSoze | because blob uses the serial port. |
| 23:41:58 | vosque | console. |
| 23:42:05 | KeyserSoze | but i shut off the tty on ttySA0 |
| 23:42:05 | vosque | Like, can you log in on the serial port? |
| 23:42:07 | cbiffle | *nod* But you're into Linux already at this point, right? |
| 23:42:13 | cbiffle | Okay. |
| 23:42:14 | KeyserSoze | yes, into linux. |
| 23:42:18 | KeyserSoze | no, i cannot log in. |
| 23:42:21 | vosque | hm.. |
| 23:42:23 | KeyserSoze | (but i can in blob) |
| 23:42:52 | cbiffle | Right. |
| 23:43:46 | cbiffle | Do you have any other machines with serial ports? |
| 23:43:50 | cbiffle | (i.e. computers? :-) |
| 23:43:56 | vosque | ya |
| 23:44:02 | vosque | spamming the serial port to recreate the problem. |
| 23:44:05 | vosque | SPAM CEREAL |
| 23:44:07 | vosque | yay |
| 23:44:28 | cbiffle | If you have a null modem cable like you use to talk to blob, try blowing a bunch of data at the port and see if it nukes the phone. |
| 23:45:17 | KeyserSoze | i just tried that, with minicom. |
| 23:45:29 | KeyserSoze | i typed and pasted a bunch of text, it did nothing. |
| 23:45:29 | cbiffle | Results? |
| 23:45:36 | cbiffle | You're in Linux? |
| 23:45:40 | cbiffle | Or some sort of Unix? |
| 23:45:51 | KeyserSoze | yes, linux |
| 23:46:06 | cbiffle | Try 'cat /dev/urandom > ---whatever port is tied to the phone---' to get more variation in the data sent. |
| 23:46:08 | KeyserSoze | i'll pipe text to it. |
| 23:46:15 | KeyserSoze | oh, good idea. |
| 23:46:24 | vosque | Have you turned the 'Reboot Computer' feature of your NMEA compatible GPS reciever off? :P |
| 23:46:42 | cbiffle | Incidentally, if you're bored, 'cat /dev/urandom > /dev/fb0' on the phone itself makes pretty snow pictures. |
| 23:47:20 | KeyserSoze | no I left that on, so that I will still be able to reboot my computer. |
| 23:47:43 | KeyserSoze | ack. it doesn't do anything. |
| 23:48:09 | cbiffle | No nothing? Hm. Curses. |
| 23:48:25 | vosque | I don't think the GPS reciever uses curses. |
| 23:48:27 | cbiffle | Your problem is very strange, grasshopper. Perhaps as_ can help us. |
| 23:48:49 | as_ | Evnin all. |
| 23:48:55 | as_ | I doubt it. |
| 23:48:58 | sjm | For what it's worth, I had this problem with an m68k board - the serial driver was looking for a ^R (0x12) then rebooting whenever it received it. Took a while to find this one... |
| 23:49:07 | cbiffle | as_, are you aware that the TuxScreen seems to have an added 'Let my GPS reboot the whole damn machine' feature? :-) |
| 23:49:23 | vosque | guessed it could be something like that. |
| 23:49:32 | cbiffle | sjm: Some older Sun machines will do that in response to a BREAK character, so that's not unheard of. Linux doesn't do that to my knowledge. |
| 23:49:33 | vosque | strikes me as the kind of thing that could happen to a tux. |
| 23:49:49 | as_ | I'm not convinced I know what GPS means in this context. |
| 23:49:58 | vosque | global positioning system |
| 23:50:12 | cbiffle | as_: The ultimate wake-up-in-the-gutter-and-figure-out-how-to-get-home device. |
| 23:50:25 | as_ | Ok, so that was right, but how does it tie into the tuxscreen? |
| 23:50:38 | vosque | someone is using one on their serial port. |
| 23:50:41 | cbiffle | sjm: Hm. On the other hand, if there was a nasty catch with a BREAK-like character, that -would- explain why we have to uuencode files passed to blob. |
| 23:50:42 | KeyserSoze | i took the text from the GPS I had in minicom, put it in a file, and sent it to the tuxscreen. it didn't reboot. |
| 23:50:59 | cbiffle | KeyserSoze: Oooh, good thinking. |
| 23:51:11 | KeyserSoze | NMEA GPS aren't *supposed* to send anything but ASCII |
| 23:51:20 | cbiffle | Right, probably no extended characters. |
| 23:51:31 | KeyserSoze | and in minicom, all I see is regular letters. |
| 23:51:37 | cbiffle | Good. |
| 23:51:44 | as_ | If the serial port had an error in reception, I think it can assert NMI, which would probably reboot in hardware. Just a guess. |
| 23:52:10 | cbiffle | *nod* Wouldn't surprise me terribly if a full UART FIFO constituted an NMI condition on this boar.d |
| 23:52:24 | KeyserSoze | doh! |
| 23:52:37 | cbiffle | doh? |
| 23:53:06 | KeyserSoze | 'doh!'. meaning, it would really suck if it rebooted when the FIFO is full. |
| 23:53:16 | cpenguin_meeting | Thats doubtful, though |
| 23:53:38 | cbiffle | *nod* But I would have initially described the idea that an NMEA GPS could reboot a computer as 'doubtful,' so I'm grasping at straws here. :-) |
| 23:53:41 | cpenguin_meeting | Because the serial code is standard amongst all ARM products |
| 23:54:07 | as_ | Having the serial port wired to the nuke russia button has always struck me as weird. |
| 23:54:22 | cbiffle | as_: Ooh. Nuke Russia, let's see, that's GPIO_16, right? |
| 23:54:49 | cpenguin_meeting | Yeah.... |
| 23:55:08 | as_ | NMI. The big unrecoverable error emergency system abort before we get curruption. |
| 23:55:11 | cbiffle | cpenguin_meeting: At the low character rate we're dealing with, do you have any thoughts on the problem? |
| 23:55:20 | cpenguin_meeting | Whats really cool, is if you plug a speaker in, it will do a little HAL "I can't do that dave" routine before it actually starts firing off missiles |
| 23:55:35 | cbiffle | as_: Yup, I've traditionally used it to signal power-fail. |
| 23:55:43 | cpenguin_meeting | cbiffle: Who's reading the data coming in from the serial port? |
| 23:55:44 | cbiffle | cpenguin_meeting: Ooh, gotta try that. Do you need the shanip board for that? |
| 23:55:51 | cbiffle | cbiffle: KeyserSoze. :-) |
| 23:56:05 | KeyserSoze | the other thing is, my GPS gets power from a PS2 plug. I have a little DC-DC converter that outputs 5 volts, and i use that to power the PS2 plug on the GPS. it works on a 80188 single board computer I have, using that as a power source. |
| 23:56:18 | as_ | Particually since the printer port doesnt have that, and I wish it did. |
| 23:56:23 | cpenguin_meeting | cbiffle: ok, this is part where you learn that I tend to personify software. What software process is reading the data from the serial port? |
| 23:56:41 | cbiffle | cpenguin_meeting: I don't know. I was suggesting that we ask KeyserSoze. :-) |
| 23:56:51 | cpenguin_meeting | KeyserSoze: Your thoughts? |
| 23:57:00 | KeyserSoze | it reboots whether my application opens the port or not. |
| 23:57:23 | KeyserSoze | or are you asking me what is reading data, when my application isn't? |
| 23:57:24 | cbiffle | KeyserSoze: What is this DC-DC converter converting -from-? |
| 23:57:25 | cpenguin_meeting | Do you have access to a X terminal? |
| 23:57:49 | cpenguin_meeting | on the tuxscreen I mean |
| 23:57:59 | KeyserSoze | cbiffle: it's not really *from* anything. it's *for* the GPS. |
| 23:58:18 | cbiffle | ... Okay. What is your source of electrons, on a large scale. The wall? A battery? |
| 23:58:26 | vosque | THE SUN |
| 23:58:29 | cbiffle | We're having fun-with-language day here in #tuxscreen. :-) |
| 23:58:52 | KeyserSoze | yes, i do have a x term. |
| 23:58:52 | KeyserSoze | i can show you the running processes if you like. |
| 23:58:52 | KeyserSoze | i forgot to mention, I am in picogui while doing this. |
| 23:59:47 | cpenguin_meeting | KeyserSoze: ok then - hook up the ol' GPS, and do a big cat /dev/ttySA0 > foo and see what happens |
