**** BEGIN LOGGING AT Mon Jun 13 23:59:56 2005
Jun 14 00:04:30 <VirtualCtor> Well, time to turn-in.  'Nite all.
Jun 14 00:18:28 <jf-work> morning
Jun 14 00:30:56 <jf-work> bob_tm : I'm loosing TAB while saving the cups.mk you sent me on disk (using IE)
Jun 14 00:31:32 <bob_tm> jf-work: Ok - it was pasted into Gmail, so I'm not really surprised. I'll resend
Jun 14 00:31:33 <jf-work> bob_tm: could you save it somewhere on the web and push me a URL which I can wget ?
Jun 14 00:31:50 <bob_tm> mom.
Jun 14 00:33:13 <bob_tm> http://folk.uio.no/ingeba/cups.mk
Jun 14 00:33:40 <jf-work> ty
Jun 14 00:33:45 <bob_tm> np
Jun 14 00:34:51 <jf-work> wiley building
Jun 14 00:38:52 <bob_tm> Native?
Jun 14 00:39:32 <jf-work> cross. I meant "I'm building for wiley"
Jun 14 00:39:38 <jf-work> ok, it's done
Jun 14 00:40:39 <bob_tm> Looking good - no "Can't find shared lib libyetanothergraphicformat.so-0.0.1" ?
Jun 14 00:43:41 <jf-work> nothing I can see flashing. Just an error (you may also have it) saying 'cannot change ownership of /opt/bin/lppassword'
Jun 14 00:43:56 <bob_tm> Nope.
Jun 14 00:44:16 <jf-work> From Makefile in cups/systemv I guess
Jun 14 00:44:24 <jf-work> at -ipk time
Jun 14 00:44:25 <bob_tm> Ah, during build 
Jun 14 00:44:33 <jf-work> are you building under root ?
Jun 14 00:44:39 <bob_tm> That is normal. We should put that in postinst.
Jun 14 00:44:46 <bob_tm> No :-o 
Jun 14 00:44:59 <jf-work> ok
Jun 14 00:45:26 <bob_tm> It will have to wait until this evening. Can you run it and get into the web server?
Jun 14 00:45:46 <bob_tm> ...or should I  commit now and we can add things later?
Jun 14 00:46:11 <jf-work> No hurry.
Jun 14 00:46:13 <bob_tm> btw. are there separate ready_for_testing variables for the targets wl500g/unslung.
Jun 14 00:46:30 <jf-work> yep
Jun 14 00:47:15 <bob_tm> Good. I will commit and put cups in wl500g ready for testing. This evening I will add the chown of lppassword to postinst (not really a big deal)
Jun 14 00:47:26 <bob_tm> Agree?
Jun 14 00:47:37 <jf-work> agreed.
Jun 14 00:47:53 <jf-work> that's cool. thank you very much
Jun 14 00:47:55 <bob_tm> Goed - have a good one JF.
Jun 14 00:48:03 <bob_tm> thx
Jun 14 00:52:19 <CIA-12> 03marceln * 10unslung/ (7 files in 4 dirs): Package add: tftp-hpa, vblade
Jun 14 00:55:29 <CIA-12> 03ingeba * 10unslung/make/cups.mk: Added jeanfabrice's wl500g mods and added native support for jpeg and tiff
Jun 14 01:00:26 <CIA-12> 03ingeba * 10unslung/Makefile: Added cups to wl500g packages ready for testing and made metalog obsolete
Jun 14 01:53:42 <marceln> Hi Does somebody know a way to reach the "Slugbug maintainer"?
Jun 14 01:58:06 <CIA-12> 03jeanfabrice * 10unslung/ (7 files in 2 dirs): Promote cups for wiley. Control file is now generated
Jun 14 02:09:40 <CIA-12> 03jeanfabrice * 10unslung/Makefile: Promote tftp-hpa, vblade for wiley
Jun 14 02:16:16 <marceln> Hi CIA-12: I need to add my name and the packages vblade and tftp-hpa to slugbug. Do you know how to do that?
Jun 14 02:22:01 <jf-work> marceln : dunno, sorry. btw doesn't tftp-hpa should depend on readline ?
Jun 14 02:25:14 <marceln> I have to check. But i think that tftp does and in.tfpd doesn't.
Jun 14 02:27:18 <jf-work> main.c:58:31: readline/readline.h: No such file or directory (while builing tftp-hpa)
Jun 14 02:28:25 <jf-work> maybe a problem from me... Let me check
Jun 14 02:28:52 <marceln> I See. The code is:
Jun 14 02:28:52 <marceln>   #ifdef WITH_READLINE
Jun 14 02:28:52 <marceln>   #include <readline/readline.h>
Jun 14 02:30:37 <marceln> It didn't use readline when i compiled it.
Jun 14 02:30:53 <marceln> So i didn't see it could use it.
Jun 14 02:32:38 <jf-work> build machine fails building - for wiley target - But it builds fine on my env.
Jun 14 02:32:59 <marceln> What's a wiley target?
Jun 14 02:34:08 <jf-work> sorry : unslung build system is also use for compiling optware (package) for Asus WL products family
Jun 14 02:34:26 <jf-work> Asus Wl = wiley
Jun 14 02:35:44 <jf-work> Well, on my own env, I don't have readline in staging dir. tftp-hpa compiles and build fine. The build machine use to have readline staged and tftp-hpa failed
Jun 14 02:38:25 <marceln> So it looks like the build fails if you have installed readline.
Jun 14 02:39:20 <jf-work> when readline is staged, yes
Jun 14 02:40:53 <dyoung> marceln, does the slugbug not let you create a new acct?
Jun 14 02:41:12 <marceln> I have an account and filed in a bug.
Jun 14 02:41:24 <marceln> Thats the first reason i came here at all.
Jun 14 02:41:52 <dyoung> ok
Jun 14 02:42:01 <marceln> I just want to packages registered and my name as developer for those packages.
Jun 14 02:42:31 <dyoung> thats defined in the control file for old-style packges, or in the Makefile for new-style packages.
Jun 14 02:42:52 <dyoung> or are you reffering to specific slug-bug assignments?
Jun 14 02:43:52 <marceln> When i tried to fill in bug 154 the package atftp wasn't on the list.
Jun 14 02:44:59 <marceln> I didn't like that so thats why i try to get the packages on the list.
Jun 14 02:45:38 <marceln> Besides that "http://www.nslu2-linux.org/wiki/Unslung/AddAPackageToUnslung" states that:
Jun 14 02:45:38 <marceln>   Send developer registration to the Slugbug maintainer.Send package registration to the Slugbug maintainer.
Jun 14 02:45:59 <dyoung> Ahh, in the compnents dropdown list?
Jun 14 02:46:04 <dyoung> Its in there now.
Jun 14 02:46:12 <marceln> Thank you.
Jun 14 02:46:28 <dyoung> I didnt do it, someone else must have. :-)
Jun 14 02:50:39 <jf-work> marceln: about the readline issue : when readline is already staged, tftp-hpa finds readline headers. But as readline is not mention in your makefile, the builds is done without adding the -I include option.
Jun 14 02:51:22 <marceln> jf-works: I have seen a flag "--without-readline" in the configure.
Jun 14 02:51:38 <marceln> jf-works: Could you test if this solves the problem.
Jun 14 02:52:30 <jf-work> I guess the should solve telling configure not to search for readline header
Jun 14 02:53:30 <jf-work> s/the/this
Jun 14 02:57:02 <jf-work> tftp-hpa builds properly with --without-readline and a staged readline
Jun 14 03:00:04 <marceln> I have changed the makefile and testing the build 2 times.
Jun 14 03:00:57 <jf-work> ok
Jun 14 03:02:38 <marceln> The cross compile build works
Jun 14 03:02:47 <jf-work> cool
Jun 14 03:03:21 <marceln> I am waiting on the native build because the NSLU2 is a lot slower.
Jun 14 03:05:43 <dyoung> cross is preffered anyways.
Jun 14 03:13:29 <CIA-12> 03marceln * 10unslung/make/tftp-hpa.mk: 
Jun 14 03:13:30 <CIA-12> Added "--without-readline" to configure because there are problems
Jun 14 03:13:30 <CIA-12> with staged readline package.
Jun 14 03:29:34 <maher> can i ask a questionabout the originalfirmware thatcomes withthe nslu2?
Jun 14 04:15:29 <rwhitby> marceln: welcome!
Jun 14 04:51:27 <marceln> rwhitby: Hi. I was away from my computor.
Jun 14 04:52:51 <rwhitby> get everything sorted out?
Jun 14 04:53:26 <marceln> I have solved the big part.
Jun 14 04:53:45 <marceln> I have uploaded both packages. But they aren't in the bug system yet.
Jun 14 04:53:56 <rwhitby> yeah, that's manual
Jun 14 04:54:15 <rwhitby> and ka6sox is probably asleep at the moment
Jun 14 05:23:23 <dr_nick> hi
Jun 14 05:23:31 <rwhitby> g'day
Jun 14 05:25:45 <dr_nick> i've got a question regarding the redboot upgrade mode. i'm running a wireless network, so the windows machine i'm going to use for upgrading my slug has a wireless connection to the slug. 
Jun 14 05:25:53 <rwhitby> yep
Jun 14 05:25:55 <rwhitby> I do the same
Jun 14 05:26:16 <rwhitby> you're using the SerComm tool?
Jun 14 05:26:17 <dr_nick> i can't find anything about this on the wiki. is it safe to flash over a wlan or should i use a cross cable and connect the slug to the machine
Jun 14 05:26:48 <dr_nick> up to now i've always used http upgrade, as i'm a 90% mac user
Jun 14 05:27:00 <dr_nick> yes, but i'm going to use sercomm
Jun 14 05:27:02 <rwhitby> I always flash over the wlan
Jun 14 05:27:07 <rwhitby> using sercomm
Jun 14 05:27:41 <rwhitby> if it fails, it will fail safe, and you can always try again, or fall back to using sercomm directly connected
Jun 14 05:27:50 <dr_nick> yep
Jun 14 05:28:24 <dr_nick> that's all, thanks for your time :-)
Jun 14 05:29:34 <rwhitby> np
Jun 14 05:30:04 <rwhitby> if it works fine, then post a reassuring message to the mailing list so that others will know
Jun 14 06:40:51 <ka6sox-zzzz> NAiL Ping
Jun 14 06:42:05 <marceln> Hi ka6sox
Jun 14 06:43:19 <ka6sox-zzzz> morning.
Jun 14 06:44:14 <marceln> Could you add "vblade" and "tftp-hpa" to slugbug
Jun 14 06:45:43 <ka6sox-zzzz> I'm not sure what you are asking...anyone can add a slugbug.
Jun 14 06:46:07 <marceln> Sorry. 
Jun 14 06:46:33 <ka6sox-zzzz> its okay I just don't understand...I'm willing to help if you need it.
Jun 14 06:46:59 <marceln> I made two new packages "vblade" and "tftp-hpa". New packages should be aded to slugbug so that people are able to fill in bug request.
Jun 14 06:47:27 <ka6sox-zzzz> got it.
Jun 14 06:47:46 <marceln> My self had problems with filling in a bug for "atftp" because it wasn't in the list of packages and i don't want others to have the same problem.
Jun 14 06:48:02 <marceln> You can add my name to them.
Jun 14 06:48:03 <ka6sox-zzzz> understood.
Jun 14 06:48:07 <ka6sox-zzzz> k
Jun 14 06:48:08 <marceln> Thanks.
Jun 14 06:48:27 <ka6sox-zzzz> your contact info is in the .mk's?
Jun 14 06:48:34 <marceln> Yep.
Jun 14 06:48:55 <ka6sox-zzzz> thats all I need!
Jun 14 06:49:02 <ka6sox-zzzz> thanks for developing!
Jun 14 07:07:32 <NAiL> ka6sox-zzzz: Pong
Jun 14 07:09:07 <ka6sox-zzzz> email bounced...but I figured out I typed it wrong....
Jun 14 07:09:16 <NAiL> oyvind@repvik.rog
Jun 14 07:09:19 <NAiL> bah, org
Jun 14 07:09:36 <ka6sox-zzzz> lysdexic like me!
Jun 14 07:19:11 <ka6sox-zzzz> emails away!
Jun 14 07:20:09 <NAiL> got it
Jun 14 07:21:48 <NAiL> was the name decided?
Jun 14 09:43:03 <CIA-12> 03bzhou * 10unslung/ (Makefile make/tsocks.mk): tsocks ready for cross testing
Jun 14 09:58:03 <ka6sox-office> NAiL, ping
Jun 14 13:07:05 <NAiL> ka6sox-office: pong
Jun 14 13:21:11 <ka6sox-office> did you get the keys?
Jun 14 13:21:19 <NAiL> yeah
Jun 14 13:21:20 <ka6sox-office> oh..thats right...
Jun 14 13:21:26 <ka6sox-office> i need the other key...
Jun 14 13:21:35 <ka6sox-office> ~slugtime
Jun 14 13:21:36 <jbot> i guess slugtime is 40 hour days 10 hours in 4 timezones with overlap
Jun 14 13:21:48 <ka6sox-office> waiting for the TZ behind me to wake up.
Jun 14 13:54:47 <NAiL> tz behind you? Aren't you in the UK?
Jun 14 14:11:17 <bob_tm> jeanfabrice: ping?
Jun 14 14:15:01 <jeanfabrice> bob_tm: pong
Jun 14 14:48:49 <ka6sox-office> NAiL, I'm left coast of the US.
Jun 14 14:49:07 <NAiL> aha
Jun 14 15:55:19 <Christopher> #handhelds.org
Jun 14 15:56:38 <dyoung> were you accepted into the google programme with handhelds.org ?
Jun 14 15:59:14 <Christopher> haven't heard yet
Jun 14 15:59:26 <Christopher> deadline for applications is tonight
Jun 14 15:59:34 <Christopher> should know for sure by the 24th
Jun 14 15:59:40 <dyoung> Ah, okay I thought you were celebrating. :-)
Jun 14 16:00:47 <Christopher> no, just trying to join the channel :-)
Jun 14 16:01:10 <Christopher> finally got reliable internet
Jun 14 16:01:28 <Christopher> last week I haven't been able to get on much
Jun 14 16:02:00 <jacques> Christopher, where are you now?
Jun 14 16:02:07 <Christopher> maryland
Jun 14 16:02:09 <rwhitby> Christopher: can we change the -U option in UpSlug to "-Do-You-Really-Know-What-You-Are-Doing" (case sensitive) ?
Jun 14 16:02:19 <Christopher> yes
Jun 14 16:02:44 <Christopher> that would be better really
Jun 14 16:02:46 <rwhitby> and not advertise it until we've tested it thouroughly
Jun 14 16:02:52 <Christopher> sounds good to me
Jun 14 16:02:56 <rwhitby> thx
Jun 14 16:03:06 <rwhitby> we'll have 100's more people using it now ...
Jun 14 16:03:18 <Christopher> ok
Jun 14 16:03:29 <Christopher> I don't have my linux machine 
Jun 14 16:03:48 <Christopher> and I haven't set up CVS on my laptop so it would be great if someone else could do it
Jun 14 16:03:58 <rwhitby> Upgrade Mode was really only put in to make your tool more popular
Jun 14 16:04:06 <Christopher> I need to dual boot my laptop, buy I haven't had the time recently
Jun 14 16:04:35 <Christopher> that's good :-)
Jun 14 16:04:58 <rwhitby> Christopher: can you ask jbowler to do it when you see him?
Jun 14 16:05:03 <Christopher> sure
Jun 14 16:06:23 <NAiL> ka6sox-office: vhost is up
Jun 14 16:06:49 <ka6sox-office> NAiL, thanks.
Jun 14 16:07:08 <ka6sox-office> okay since you have the keys I'm going to test now.
Jun 14 16:08:10 <NAiL> k
Jun 14 16:41:59 <Christopher_> jbowler-zzz: ping
Jun 14 17:36:05 <CIA-12> 03bzhou * 10unslung/make/tsocks.mk: fixed a make install problem, strip so
Jun 14 17:47:51 <rwhitby-away> almost 300 unslung 5.5 downloads, and only one error reported (which looks like a Linksys bug)
Jun 14 17:48:38 <jacques> that disk2 thing?
Jun 14 17:48:40 <NAiL> :)
Jun 14 17:48:52 <rwhitby-away> yeah, has anyone here confirmed that?
Jun 14 17:49:21 <jacques> I haven't tried
Jun 14 17:50:00 * rwhitby-away does a "mv /initrd/.sda2root /initrd/.sdb2root" and tries it ....
Jun 14 17:53:16 <rwhitby-away> Disk 2/Flash: Ready, 0MB (0% Free) 
Jun 14 17:53:21 <rwhitby-away> confirmed
Jun 14 17:54:24 <jacques> darn you linksys
Jun 14 17:57:45 <rwhitby-away> index.cgi is what prints it, and that seems to call stuff from libhd.so
Jun 14 17:57:52 <rwhitby-away> (none of which we have source for)
Jun 14 17:59:28 <jacques> again I say, darn you linksys
Jun 14 17:59:59 <rwhitby-away> hmm - /proc/partitions is empty
Jun 14 18:00:30 <rwhitby-away> is it empty on a normal slug with disk in /dev/sda?  (saves me rebooting if someone else can answer)
Jun 14 18:01:52 <rwhitby-away> anyone?
Jun 14 18:02:20 <jacques> I don;t have any unslung slugs right now
Jun 14 18:11:01 <rwhitby-away> # cat /proc/partitions
Jun 14 18:11:01 <rwhitby-away> major minor  #blocks  name
Jun 14 18:11:01 <rwhitby-away>    8     0    1429344 scsi/host1/bus0/target0/lun0/disc
Jun 14 18:11:01 <rwhitby-away>    8     1    1237415 scsi/host1/bus0/target0/lun0/part1
Jun 14 18:11:02 <rwhitby-away>    8     2     127673 scsi/host1/bus0/target0/lun0/part2
Jun 14 18:11:02 <rwhitby-away>    8     3      63135 scsi/host1/bus0/target0/lun0/part3
Jun 14 18:11:04 <rwhitby-away> #
Jun 14 18:11:44 <rwhitby-away> so those are the contents when the disk is in disk1
Jun 14 18:11:51 <rwhitby-away> and it's empty when the disk is in disk2
Jun 14 18:13:28 <rwhitby-away> wonder what it does when rootfs is jffs2, and disk is plugged into disk2 ...
Jun 14 18:14:45 <glc> /proc/partitions empty.
Jun 14 18:16:25 <rwhitby-away> # df
Jun 14 18:16:25 <rwhitby-away> Filesystem           1k-blocks      Used Available Use% Mounted on
Jun 14 18:16:25 <rwhitby-away> rootfs                    6528      5308      1220  81% /
Jun 14 18:16:25 <rwhitby-away> "/dev/mtdblock4            6528      5308      1220  81% /
Jun 14 18:16:26 <rwhitby-away> "/dev/mtdblock4            6528      5308      1220  81% /dev.state
Jun 14 18:16:26 <rwhitby-away> "/dev/mtdblock4            6528      5308      1220  81% /var.state
Jun 14 18:16:28 <rwhitby-away> "/dev/sdb1              1217924     32864   1172688   3% /share/flash/data
Jun 14 18:16:30 <rwhitby-away> "/dev/sdb2               123627     65838     56513  54% /share/flash/conf
Jun 14 18:16:32 <rwhitby-away> # cat /proc/partitions
Jun 14 18:16:34 <rwhitby-away> major minor  #blocks  name
Jun 14 18:16:36 <rwhitby-away> #
Jun 14 18:16:45 <rwhitby-away> ok, so it's nothing to do with rootfs location, just disk2.
Jun 14 18:17:17 * rwhitby-away double checks with jffs2 and disk1, just to be sure
Jun 14 18:19:29 <rwhitby-away> I wonder what /proc/partitions contains when disk2 is used with stock R29 ...
Jun 14 18:20:34 <rwhitby-away> back later ..
Jun 14 19:03:58 <jbowler-zzz> rwhitby-away, christopher: upslug -U current outputs "upslug: -U: disabled" and promptly does an exit(1).  If anyone wants to use it they have to edit the source - at which point they will see the comment "For safety this is disabled - it will overwrite redboot."
Jun 14 19:08:51 <christopher> jbowler, that seems good enough to me
Jun 14 19:11:10 <jbowler-zzz> I think so, most people are building from soure themselves anyway - I put upslug-native in as a package but only openslug builds it.
Jun 14 19:11:42 <jbowler-zzz> Did you have any thoughts about the license to use?
Jun 14 19:11:54 <christopher> jbowler-zzz, yeah anyone who is editing the source is already taking it into their own hands
Jun 14 19:12:16 <christopher> yes, I'm thinking gpl that way we can make sure the source is distributed
Jun 14 19:12:56 <christopher> you wrote parts of it, so what are you thoughts on it?
Jun 14 19:13:03 <jbowler-zzz> Ok - would you like me to add the GPL to the CVS?  (I.e. add the COPYING file etc.)
Jun 14 19:13:56 <christopher> sounds good to me 
Jun 14 19:14:12 <jbowler-zzz> I'm fine with GPL - I don't like it myself, but it works fine for this and I'm happy working with it.
Jun 14 19:15:55 <christopher> alright, we'll go with that then if you're fine with it
Jun 14 19:16:38 <christopher> I've never really investigated licenscing a whole lit
Jun 14 19:16:42 <christopher> s /lit/lot
Jun 14 19:17:06 <christopher> anyway, I got linux up and running again so maybe I can be of a little more help around here :-)
Jun 14 19:17:41 <jbowler-zzz> I'll get the relevant file from www.opensource.org - http://www.opensource.org/licenses/gpl-license.php - and put it in the 'standard' place (i.e. where autoconf would look for it.)
Jun 14 19:19:31 <rwhitby-away> jbowler-zzz: excellent
Jun 14 19:20:58 <christopher> jbowler-zzz: thanks
Jun 14 19:21:19 <rwhitby-away> peteru-home: welcome back
Jun 14 19:26:19 <peteru-home> rwhitby: Hello there. Just got home :-)
Jun 14 19:26:27 <rwhitby> good trip?
Jun 14 19:26:59 <peteru-home> yeah!
Jun 14 19:27:51 <peteru-home> I'm stuffed though - spent the last 30+ hours on airplanes.
Jun 14 19:28:24 <peteru-home> Santiago -> Buenos Aires -> Auckland -> Sydney
Jun 14 19:29:25 <peteru-home> any major developments in the slug world over the last 6-8 weeks?
Jun 14 19:31:19 <rwhitby> beta of openslug, and two major releases of unslung
Jun 14 19:31:26 <rwhitby> nothing much :-)
Jun 14 19:32:13 <rwhitby> jbowler is now co-lead of OpenSlug development with [g2] :-)
Jun 14 19:32:25 * [g2] hugs jbowler-zzz 
Jun 14 19:34:08 <[g2]> rwhitby, did yo catch the news in openjtag about APEX ?
Jun 14 19:34:41 <jacques> [g2], what's the latest?
Jun 14 19:34:55 <rwhitby> [g2]: yep saw that
Jun 14 19:35:24 <[g2]> beewoolie and I worked out the latest bugs and I booted from the zImage in /boot on the jffs2
Jun 14 19:35:45 <jacques> rock on!
Jun 14 19:36:25 <[g2]> I even zero'd out memory where the kernel was loaded and then watched it get filled in properly from the load from jffs2
Jun 14 19:37:08 <[g2]> with APEX an ipkg upgrade of the kernel is all that'd be required on the rootfs
Jun 14 19:37:31 <[g2]> and copied to the jffs2 
Jun 14 19:37:49 <[g2]> until we put booting directly off the external device in APEX
Jun 14 19:39:04 <jacques> that will rock ever harder
Jun 14 19:39:27 <[g2]> I don't think there's a huge rush for that
Jun 14 19:39:55 <[g2]> dual rootfs's tho would rock
Jun 14 19:39:58 <NAiL> would be neat, but it isn't exactly terribly useful, it it?
Jun 14 19:40:17 <[g2]> NAiL, the boot to external ?
Jun 14 19:40:30 <NAiL> yeah
Jun 14 19:40:40 <[g2]> Oh.. I think that's very useful
Jun 14 19:40:57 <rwhitby> ok, so where is the wiki page which says where the source repo is for apex?
Jun 14 19:41:16 <rwhitby> seems like it's ready for some more developers to get on board the testing train
Jun 14 19:41:31 <rwhitby> maybe we need an APEX package in OE :-)
Jun 14 19:41:39 <[g2]> nod
Jun 14 19:41:53 <[g2]> I've got my recipie all ready to go
Jun 14 19:42:16 <NAiL> [g2]: Ok, on second thought, it might actually be useful :)
Jun 14 19:42:37 <[g2]> sure .. failsafe in JFFS2 and just updates to the external
Jun 14 19:42:43 <NAiL> yeah
Jun 14 19:42:55 <[g2]> if the external ever fails, then boot off the internal
Jun 14 19:43:09 <[g2]> there's all kinda stuff we'll be able to do
Jun 14 19:43:29 <[g2]> http://wiki.buici.com/twiki/bin/view/Main/ApexBootloader#IXP42x_and_the_Linksys_NSLU2_aka
Jun 14 19:43:38 <[g2]> ftp://ftp.buici.com/pub/apex
Jun 14 19:43:40 <NAiL> But then you'd have to support multiple fs... won't that take up a fair bit of space?
Jun 14 19:43:54 <[g2]> ext2 and jffs2 are still < 64k
Jun 14 19:44:27 <[g2]> rwhitby, you'll want the 1.2.16 release
Jun 14 19:44:46 <[g2]> I tested with 1.2.15 but there was a bug and ext2 needs to be enabled
Jun 14 19:44:54 <[g2]> I'll put up a wiki page
Jun 14 19:45:22 <rwhitby> now that 5.5 is out, I have a spare slug with a jtag port on it ...
Jun 14 19:45:41 <[g2]> cool
Jun 14 19:46:07 <rwhitby> I might even have to use that digilent cable one day, but I expect to be on the safe side of the APEX curve by now ...
Jun 14 19:46:20 <NAiL> rwhitby: one mirror is up btw... Just waiting for it to be filled now :)
Jun 14 19:46:36 <[g2]> you know the funny thing is I *still* don't have JTAG on a slog
Jun 14 19:46:38 <[g2]> slug
Jun 14 19:46:55 <[g2]> I've got a hookup on the fatslug but I haven't fully debugged it
Jun 14 19:46:59 <rwhitby> NAiL: yeah, thanks for contributing your pipe and space
Jun 14 19:47:16 <NAiL> np. It's just sitting there
Jun 14 19:47:27 <[g2]> I do have the JTAG on the avila board which was very helpful
Jun 14 20:18:44 <peteru-home> what was the eventual fall out from bk pulling the plug on open source users? I left civilisation just after the issue popped up...
Jun 14 20:22:28 <dyoung-web> peteru-home: Its still falling out.
Jun 14 20:22:41 <[g2]> rwhitby, http://www.nslu2-linux.org/wiki/HowTo/RunApex
Jun 14 20:23:01 <dyoung-web> I'm not sure that any development group has any firm direction they with so move in for SCM.
Jun 14 20:23:33 <dyoung-web> s/with so/wish to/
Jun 14 20:23:54 <dyoung-web> that was so dyslexic of me.
Jun 14 20:27:11 <peteru-home> bummer, I was expecting at least one or two alternatives to emerge by now.
Jun 14 20:27:42 <rwhitby> [g2]: do you have an apex.bb or should I make one?
Jun 14 20:28:02 <[g2]> I don't have one now
Jun 14 20:28:44 <[g2]> rwhitby, I'd hold off on putting it in OE for now
Jun 14 20:29:01 <rwhitby> peteru-home: but we do have an svn repo set up on berlios.de as a backup plan in case OE can't decide :-)
Jun 14 20:29:22 <rwhitby> [g2]: why? you know that I don't build anything unless the process is completely automated ....
Jun 14 20:29:52 <[g2]> rwhitby, well that's partially the point
Jun 14 20:30:10 <[g2]> the default nslu2_config isn't fully correct yet
Jun 14 20:30:46 <[g2]> also I'd prefer not to be fielding questions from Redboot reflashers unless they know what they are doing
Jun 14 20:31:12 <[g2]> I don't think waiting a few days is holding anyone up
Jun 14 20:31:38 <rwhitby> [g2]: cool, I'll start on one but not push it
Jun 14 20:32:03 <[g2]> ok, if you want
Jun 14 20:32:08 <[g2]> it'll be trivial to add
Jun 14 20:32:42 <peteru-home> is apex much smaller than RB?
Jun 14 20:32:59 <dyoung-web> Its TINY.
Jun 14 20:32:59 <[g2]> peteru-home, < 64K 
Jun 14 20:33:18 <[g2]> it quadrupled with ext2 and jffs2
Jun 14 20:33:41 <peteru-home> excellent! 1 erase block including jffs2?
Jun 14 20:34:05 <[g2]> yes with room to grow :)
Jun 14 20:34:37 * peteru-home gives out a hand full of chocolates to apex hackers.
Jun 14 20:34:39 <[g2]> plus with OpenSlug we can now regain 1M + 128K*3 if we want
Jun 14 20:35:36 <[g2]> APEX has a config block which will be the second block
Jun 14 20:36:01 <[g2]> I haven't written to it or tested it but it's there
Jun 14 20:36:33 <[g2]> so Syconf, the Ramdisk block, and the trailer block would also be freed
Jun 14 20:36:42 <dyoung-web> are we gonna hide a mtd part table someplace or specify that on the cmdline? 
Jun 14 20:37:20 <rwhitby> apex should pass that to the kernel in a behind-the-scenes cmdline
Jun 14 20:37:23 <[g2]> dyoung-web, I'd think we could move it to inside the APEX config block
Jun 14 20:37:49 <[g2]> APEX already fully passes the TAGS to the kernel
Jun 14 20:37:58 <[g2]> you can just change them on the boot command
Jun 14 20:38:08 <jacques> wow, just like a real bootloader
Jun 14 20:38:28 <[g2]> jacques, you don't read the arm-linux ml ?
Jun 14 20:38:30 <rwhitby> cool, so apex's knowledge of the flash layout is automatically passed to the kernel and used, and we don't have to worry about it ever again
Jun 14 20:38:49 <[g2]> rwhitby, not quite that cool, but it could be
Jun 14 20:39:02 <rwhitby> (and openslug and unslung already are both configured to take note of partition tables on the cmdline
Jun 14 20:39:05 <[g2]> right now APEX doesn't read the partition table
Jun 14 20:39:06 <jacques> [g2], not lately
Jun 14 20:39:17 <[g2]> it's hard wired for the area
Jun 14 20:39:22 <rwhitby> [g2]: with apex, there wouldn't need to be a partition table
Jun 14 20:39:40 <rwhitby> apex passes the info to the kernel, and the "partition table" only exists in memory data structures
Jun 14 20:40:42 <dyoung-web> right., that was my rudimentry understanding.
Jun 14 20:40:44 <[g2]> right now if we toasted the partition table in the last block I think it would fail
Jun 14 20:40:51 <rwhitby> Hmm - I just thought of something - Unslung using RedBoot Upgrade Mode has just introduced a bootloader dependency :-(
Jun 14 20:41:06 <dyoung-web> Yes.
Jun 14 20:41:12 <dyoung-web> we talked about it before.
Jun 14 20:41:18 <[g2]> many times
Jun 14 20:41:20 <rwhitby> (cause 5.x just clears the last block to indicate a desire to enter upgrade mode)
Jun 14 20:42:05 <dyoung-web> I dont think Joe Unslinger will need to run alternate bootloaders on their nslu2.
Jun 14 20:42:09 <rwhitby> so the question is, do we still reserve the last block for partition table and apex config info, and therefore have everything except the first and last block free for jfffs2?
Jun 14 20:42:25 <rwhitby> dyoung-web: maybe not, but the developers sure will
Jun 14 20:42:49 <[g2]> rwhitby, Ok the first two block are for APEX and it's config block
Jun 14 20:43:16 <[g2]> I'd relocate or punt the partition table and we can use the TAGS on the cmdline after they've been testing
Jun 14 20:43:18 <[g2]> I'd relocate or punt the partition table and we can use the TAGS on the cmdline after they've been tested
Jun 14 20:43:24 <rwhitby> [g2]: does it have to be first two blocks - can it be one at the start and one at the end?  or is the apex code larger than one block now?
Jun 14 20:44:10 <[g2]> ls -l src/arch-arm/rom/apex.bin
Jun 14 20:44:10 <[g2]> -rwxr-xr-x  1 tom users 47272 Jun 14 18:39 src/arch-arm/rom/apex.bi
Jun 14 20:44:21 <jacques> heh
Jun 14 20:44:31 <jacques> it's petite
Jun 14 20:44:38 <[g2]> DOH!
Jun 14 20:44:44 <[g2]> that the 14 version
Jun 14 20:44:45 <dyoung-web> thats in the realm of RFS.
Jun 14 20:45:13 <dyoung-web> but not as small as the 2k spinner program.
Jun 14 20:45:14 <dyoung-web> heh
Jun 14 20:45:22 <[g2]> ls -l src/arch-arm/rom/apex.bin
Jun 14 20:45:22 <[g2]> -rwxr-xr-x  1 tom users 51364 Jun 14 21:10 src/arch-arm/rom/apex.bin
Jun 14 20:45:40 <[g2]> it has a spinner :)
Jun 14 20:45:53 <dyoung-web> I like the spinner.
Jun 14 20:46:08 <dyoung-web> its my favorite part. :-)
Jun 14 20:46:40 <dyoung-web> "Apex not only lets you boot custom memory sizes, It adds a level of 'Bling' to your hardware!"
Jun 14 20:46:46 <jacques> every app should have at least one spinner
Jun 14 20:47:06 <rwhitby> ok, so we (well, probably me, cause Bewoolie and [g2] rightly won't care about Unslung) need to work out how to use apex with both unslung and openslug
Jun 14 20:47:53 <[g2]> rwhitby, APEX works perfectly with openslug right now
Jun 14 20:47:54 <dyoung-web> it will work, you just wont be able to upgrade it using the sercomm upgade mode.
Jun 14 20:48:14 <dyoung-web> so worse to worse, you 'll have to load it serially.
Jun 14 20:48:16 <rwhitby> [g2]: yeah, I know.  I need to work out how to make it work with both
Jun 14 20:48:36 <[g2]> I think it'll work with both right now
Jun 14 20:48:56 <[g2]> the issue is allowing certain custom configs with Openslug
Jun 14 20:49:04 <rwhitby> [g2]: you're missing the point.  I'm talking about when you've regained the 128Kbx3
Jun 14 20:49:26 <rwhitby> OpenSlug doesn't care about what /dev/mtd looks like
Jun 14 20:49:43 <[g2]> no, but the kernel sure does
Jun 14 20:49:44 <rwhitby> Unslung currently does (cause I inadvertently added a bootloader dependency to it)
Jun 14 20:49:57 <rwhitby> [g2]: the kernel will use whatever you tell it
Jun 14 20:50:02 <[g2]> nod.
Jun 14 20:50:15 <rwhitby> that's not what I mean by "doesn't care about"
Jun 14 20:50:19 <[g2]> but we do either have to tell it or let it find out on its own
Jun 14 20:50:33 <rwhitby> ok, let me try and make myself clear.
Jun 14 20:50:51 <[g2]> rwhitby, I think we're in full agreement
Jun 14 20:51:03 <rwhitby> Unslung 5.x has code in the web interface "Enter Upgrade Mode" code which does a write to a hard-code /dev/mtdblock5.
Jun 14 20:51:12 <rwhitby> THAT is what I'm talking about.
Jun 14 20:51:17 <[g2]> I know
Jun 14 20:51:37 <rwhitby> for OpenSlug, you can change the partition around as much as you like, and as long as you tell the kernel what you have done, then nobody cares.
Jun 14 20:51:50 <[g2]> nod
Jun 14 20:51:59 <[g2]> Openslug is more flexible
Jun 14 20:52:05 <rwhitby> That is why the issue is not with OpenSlug, it is with Unslung.
Jun 14 20:52:58 <[g2]> well a long time ago we said that Unslung Camp #1 and #2 won't be replacing bootloaders
Jun 14 20:53:21 <rwhitby> and that is why I said "we (well, probably me, cause Bewoolie and [g2] rightly won't care about Unslung) need to work out how to use apex with both unslung and openslug".  It works fine with OpenSlug.  I need to work out how to make it work fine with Unslung as well.  That will either mean convincing you and beewoolie to keep the partition table in the last block, or releasing Unslung 6.x
Jun 14 20:53:39 <rwhitby> [g2]: I say again, I am thinking of the developers of Unslung, not the users.
Jun 14 20:54:23 <[g2]> ok lets talk about the developers of Unslung for a minute
Jun 14 20:54:37 <rwhitby> As a developer who will be doing stuff in Apex, OpenSlug, and Unslung, I want to be able to easily move between OpenSlug and Unslung without having to change my bootloader each time.  And I want to be able to do that no matter whether I am currently running RedBoot or Apex.
Jun 14 20:55:01 <rwhitby> And I want to do it in a way which doesn't waste space for the Apex/OpenSlug combination.
Jun 14 20:55:19 <rwhitby> *that*'s the problem that I have.
Jun 14 20:55:21 <[g2]> well switching bootloader is trivial right now
Jun 14 20:55:48 <rwhitby> it may be trivial, but I don't want to do it every time I flash new firmware.
Jun 14 20:55:51 <[g2]> I have a copy of the redboot on the external device and it's a simple dd if=redboot.bin of=/dev/mdtblock0
Jun 14 20:56:45 <[g2]> then you'd need to add ethernet and the sercom upgrade utility to APEX to make it compatible functionailty wise
Jun 14 20:57:11 <[g2]> otherwise, you can't really test that portion 
Jun 14 20:57:40 <rwhitby> I consider the upgrade process to be separate from testing the firmware
Jun 14 20:58:26 <[g2]> ok.. so I think you just need to possibly edit the config info for APEX to allow it to reboot
Jun 14 20:58:28 <rwhitby> the means by which you get the kernel and rootfs onto the device should be orthogonal to the functionality of the kernel and rootfs you are testing
Jun 14 20:59:15 <rwhitby> the trouble is that the web code that I added to Unslung does it in a bootloader-specific way, without testing first.
Jun 14 20:59:26 <rwhitby> and therefore violates that principle
Jun 14 20:59:59 <[g2]> nod.
Jun 14 21:00:47 <[g2]> but back to unslung developers
Jun 14 21:00:54 <rwhitby> that's why I said "well, probably me, cause Bewoolie and [g2] rightly won't care about Unslung"
Jun 14 21:01:03 <rwhitby> it's not an apex problem, it's an unslung problem.
Jun 14 21:01:07 <[g2]> nod
Jun 14 21:01:17 <rwhitby> but I'm investigating whether apex can help solve the unslung problem
Jun 14 21:01:59 <rwhitby> remember that we always said that the firmware should not depend on the bootloader, but we should allow the developers to use the latest bootloader with both unslung and openslug firmware
Jun 14 21:02:28 <[g2]> right and you didn't violate that rule
Jun 14 21:02:36 <[g2]> except a little
Jun 14 21:02:42 <rwhitby> yeah, a little
Jun 14 21:02:45 <[g2]> :)
Jun 14 21:02:55 <[g2]> I'm sure it's only temporary
Jun 14 21:03:08 <rwhitby> now, if apex doesn't really care whether the config block goes in block #2 or the last block, then I can see a way forward ...
Jun 14 21:03:29 <jacques> I don't see why it should care
Jun 14 21:03:42 <[g2]> APEX doesn't care at all
Jun 14 21:03:49 <rwhitby> and we shouldn't care, cause we get the same space for jffs2 in both cases
Jun 14 21:04:23 <rwhitby> has the format of the apex config block been determined yet?
Jun 14 21:04:42 <[g2]> yeah it's what ever APEX uses
Jun 14 21:04:58 <[g2]> remember APEX was written for other ARM targets first
Jun 14 21:05:08 <rwhitby> but has that part of apex been written yet?
Jun 14 21:05:13 <[g2]> yes
Jun 14 21:05:19 <rwhitby> ah, ok.
Jun 14 21:05:25 <[g2]> it was a port to the IXP4xx
Jun 14 21:05:35 <[g2]> then some extra
Jun 14 21:05:52 * rwhitby will need to look at the apex code to see if there is a nefarious way around this problem ....
Jun 14 21:06:06 <[g2]> the IXP4xx currently doesn't use that block, but I haven't tried saving any environment variables
Jun 14 21:06:28 <[g2]> I don't think it's much of a problem at all to APEX
Jun 14 21:06:40 <rwhitby> what?
Jun 14 21:06:52 <rwhitby> (what's isn't much of a problem to apex)?
Jun 14 21:07:02 <[g2]> bewoolie already said that a 128K config block was *WAY* too much
Jun 14 21:07:32 <rwhitby> so will apex be able to use the spare space in the last block after the partition table but before the sercomm trailer?
Jun 14 21:07:42 <rwhitby> (about 124k's worth)
Jun 14 21:08:23 <[g2]> I think APEX could use that space with some minor mods
Jun 14 21:08:33 <[g2]> it might even be just a config option
Jun 14 21:08:44 <[g2]> nor:0x7E0000
Jun 14 21:08:47 <rwhitby> and can we also use that space and format for saving the openslug config info, and be able to read both the apex config and openslug config from inside openslug or unslung userspace?
Jun 14 21:09:22 <[g2]> devio already does that for openslug
Jun 14 21:09:30 <rwhitby> and therefore be able to switch between any bootloader and firmware combination without loosing any config info?
Jun 14 21:09:46 <rwhitby> [g2]: yes, I'm talking about getting devio to read/write the same area as apex and share info
Jun 14 21:10:02 <rwhitby> no need to write the mac address twice in the firmware
Jun 14 21:10:16 <[g2]> I don't know exactly where jbowler is at with all that stuff (the perserve/restore turnup functionality)
Jun 14 21:10:16 <rwhitby> ... in the flash I mean
Jun 14 21:10:36 <rwhitby> right, I'm talking future direction, not current implementation
Jun 14 21:10:38 <[g2]> to me it's pretty straight forward once we decide what to save/restore
Jun 14 21:10:54 <rwhitby> we just need to decide what we are going to do - the implementation is the easy part
Jun 14 21:10:56 <[g2]> I mean the entire sysconf area was what 3K ?
Jun 14 21:12:24 <[g2]> there are two very important features this enables
Jun 14 21:12:31 <[g2]> 1) multiple rootfs's
Jun 14 21:12:36 <rwhitby> ok, I've planted the seeds now.  The direction I'm heading is seamless transitions from any combination of bootloaders and firmware without loosing config information
Jun 14 21:12:53 <[g2]> 2) an ipkg upgrade the jffs2 of the kernel files is an upgrade of the kernel
Jun 14 21:12:58 <[g2]> on the next boot
Jun 14 21:13:27 <rwhitby> openslug devio should be able to set the next boot option directly
Jun 14 21:13:41 <rwhitby> by writing to the shared apex/openslug config area
Jun 14 21:14:09 <[g2]> well optional openslug config area
Jun 14 21:14:20 <rwhitby> why optional?
Jun 14 21:14:25 <rwhitby> it'll be there for redboot too
Jun 14 21:14:34 <rwhitby> redboot just won't use it
Jun 14 21:14:34 <[g2]> well it's not needed
Jun 14 21:14:50 <rwhitby> what's not needed by whom?
Jun 14 21:15:19 <[g2]> currently the config area is an option
Jun 14 21:15:30 <rwhitby> which config area?  apex or openslug?
Jun 14 21:15:39 <[g2]> APEX
Jun 14 21:16:13 <rwhitby> right, but I though we decided it was need for all sorts of stuff in the future (like determining which of the two rootfs candidates to boot from)
Jun 14 21:16:28 <rwhitby> remember, we are talking future direction, not current implementation
Jun 14 21:17:12 <[g2]> I'd like to see the option of having a persistent config area for the APEX bootloader avialable
Jun 14 21:17:27 <[g2]> which could be shared with others
Jun 14 21:17:49 <jacques> shared with other what?
Jun 14 21:17:59 <[g2]> openslug/unslung config
Jun 14 21:18:22 <rwhitby> agreed.  lets work towards that, and see if it can be put in the last block between the partition table and the sercomm trailer, so that it can even be preserved when you load redboot and unslung onto that slug temporarily
Jun 14 21:18:24 <jacques> shared with other dists? ok
Jun 14 21:18:54 <rwhitby> [g2]: does that sound like a worthwhile goal?
Jun 14 21:19:54 <[g2]> sure as long as beewoolie is ok with it. essentially it's only use X K of the config block and perserve the rest on erases and updates
Jun 14 21:20:17 <rwhitby> yep.  offset plus length, like the other things in apex
Jun 14 21:20:23 <[g2]> nod.
Jun 14 21:21:01 <rwhitby-treo> nod
Jun 14 21:21:21 <[g2]> it'd be something like nor:0x7E0000@1k+8k
Jun 14 21:22:07 <[g2]> that basically says used the 8k range 1K past the last block
Jun 14 21:22:27 <rwhitby-treo> g2: that sounds perfect
Jun 14 21:22:32 <[g2]> beewoolie is adding some extra stuff in like that
Jun 14 21:23:19 <[g2]> and the only other option for APEX would be to have a switch that says "perserve the other data in the config block"
Jun 14 21:23:56 <[g2]> another option would be to save stuff in memory on the warm boot
Jun 14 21:24:01 <rwhitby-treo> which he wanted to do anyway
Jun 14 21:24:23 <[g2]> beewoolie is a great guys :)
Jun 14 21:24:31 <rwhitby-treo> the preserve option that is
Jun 14 21:24:31 <[g2]> beewoolie is a great guy :)
Jun 14 21:25:12 <[g2]> preserve is really getting simple for Openslug
Jun 14 21:25:25 <rwhitby-treo> so that the config could go after the code in the same block
Jun 14 21:25:59 <[g2]> on openslug it really doesn't need to be saved actually
Jun 14 21:27:02 <[g2]> becuase with the booting from kernel in jffs2 everything could be an ipkg upgrade
Jun 14 21:27:47 <[g2]> in the case of an externally running program, the jffs2 would need to be mounted can the kernel files / modules copied over
Jun 14 21:28:11 <[g2]> on jffs2 that happens in-situ
Jun 14 21:28:17 <rwhitby-treo> cool
Jun 14 21:28:31 <[g2]> or however one spells that
Jun 14 21:29:14 <[g2]> and we definitly need to verify the serial download (which didnt' fully work for me tonight) on APEX
Jun 14 21:30:22 <[g2]> because the APEX xmode recovery is the only option other than JTAG if the kernel gets hosed and with don't have multiple rootfses in place
Jun 14 21:32:21 <[g2]> that's one of the real reasons booting directly from the ext device would be nice
Jun 14 21:32:31 <[g2]> you'd never have to change the jffs2
Jun 14 21:32:55 <[g2]> if botched the kernel, just switch the external device
Jun 14 21:33:54 * [g2] has the other thought of teaching APEX about cramfs and putting the failsafe kernel in the RO cramfs
Jun 14 21:34:30 <rwhitby-treo> like openwrt
Jun 14 21:34:50 <[g2]> I think there doing something like that
Jun 14 21:35:10 <[g2]> I think they were also playing with live kernel replacment
Jun 14 21:35:18 <[g2]> in memory
Jun 14 21:35:32 <jacques> too bad the slug only has one button
Jun 14 21:35:42 <jacques> otherwise we could bootstrap it
Jun 14 21:35:45 <[g2]> it has two
Jun 14 21:35:52 <[g2]> reset and power
Jun 14 21:36:17 <CIA-12> 03bzhou * 10unslung/ (4 files in 2 dirs): upstream upgrade to 20050607, patch configure.ac instead of configure
Jun 14 21:36:18 <jacques> oh yeah, but is reset interceptable ?
Jun 14 21:36:25 <[g2]> sure
Jun 14 21:36:28 <jacques> or does it do something hardwired
Jun 14 21:36:39 <[g2]> no it's just a GPIO
Jun 14 21:36:44 <[g2]> GPI
Jun 14 21:37:10 <jacques> then we can input the kernel code by pressing buttons, like they did in the old days
Jun 14 21:37:34 <[g2]> a long time ago beewoolie had talked about being able to select different options via presses on the power button after holding the reset 
Jun 14 21:37:35 <jacques> tho I guess they had switches and did 8 bits at a time
Jun 14 21:37:36 <rwhitby-treo> heh
Jun 14 21:37:58 <[g2]> we can even give feed back via the disk leds
Jun 14 21:38:19 <rwhitby-treo> I'd rather not use any buttons - slugs are too high to reach
Jun 14 21:38:26 <jacques> heh
Jun 14 21:38:41 <jacques> ok, here's what we do
Jun 14 21:38:53 <[g2]> rwhitby-treo, I know you don't flash bad mojo to those slugs
Jun 14 21:38:56 <jacques> build gadget and OTG into APEX
Jun 14 21:39:01 <jacques> OtG
Jun 14 21:39:15 <[g2]> hw doesn't support OTG
Jun 14 21:39:21 <jacques> dammit
Jun 14 21:39:24 <[g2]> On-the-Go
Jun 14 21:39:24 <jacques> you sure?
Jun 14 21:39:36 <rwhitby-treo> yep
Jun 14 21:39:42 <jacques> sigh
Jun 14 21:40:01 <[g2]> usb-serial is plenty
Jun 14 21:40:30 <[g2]> then just plug in a usb-serial adapter and hello bootloader if enabled
Jun 14 21:41:21 <jacques> but then you need the adaptor
Jun 14 21:41:32 <[g2]> a usb-serial adapter ?
Jun 14 21:41:51 <jacques> of course, a A-A cable is probably harder to come by
Jun 14 21:42:28 <[g2]> no the other end plugs into a serial port or a null modem and another same adapter to the PC
Jun 14 21:42:29 <jacques> or whatever cable one uses for OtG
Jun 14 21:42:31 <rwhitby-treo> telnet
Jun 14 21:43:09 <rwhitby-treo> ssh into apex
Jun 14 21:43:12 <[g2]> APEX would have to learn usb-ethernet AND an IP stack
Jun 14 21:43:19 <[g2]> and ssh
Jun 14 21:43:44 <jacques> i think ssh would be too much code
Jun 14 21:43:59 <rwhitby-treo> there's 80kb free right?
Jun 14 21:44:21 <jacques> I'm not sure we want to use the entire free for this one feature
Jun 14 21:44:25 <[g2]> actually I think the cramfs makes much more sense
Jun 14 21:44:25 <jacques> even if it would fit
Jun 14 21:44:42 <rwhitby-treo> beewoolie can do anything in 80kb
Jun 14 21:44:54 <[g2]> plus you arm the DOG and let the other rootfs chain up the DOG
Jun 14 21:45:04 <[g2]> otherwise recover
Jun 14 21:45:36 <jacques> so you have to plug in a flash drive
Jun 14 21:45:42 <[g2]> no
Jun 14 21:45:43 <jacques> with the recovery image on it
Jun 14 21:45:47 <[g2]> no
Jun 14 21:45:50 <jacques> no what?
Jun 14 21:46:19 <[g2]> no there's plenty room for a cramfs or two full jffs2 root partitions in flash
Jun 14 21:46:38 <[g2]> one cramfs and one jffs2
Jun 14 21:46:43 <[g2]> or two jffs2
Jun 14 21:46:47 <jacques> I wasn't responding to what you said
Jun 14 21:47:00 <jacques> I'm trying to think of ways to get arbitrary data to apex
Jun 14 21:47:02 <[g2]> sory
Jun 14 21:47:41 <[g2]> well right now APEX can a) read serial, b) look in memory c) read flash
Jun 14 21:48:03 <[g2]> d) get some tiny input from the push buttons
Jun 14 21:48:24 <[g2]> I don't think anybody wants to toggle in APEX via the buttons :)
Jun 14 21:48:44 <[g2]> but given a PIC and an accuator it might owrk
Jun 14 21:48:46 <[g2]> work
Jun 14 21:50:31 <jacques> too bad slug doesn't have a microphone
Jun 14 21:51:03 <[g2]> well a USB-Sound blaster would work :)
Jun 14 21:52:01 <jacques> then if we put support for that in apex, we can load data with sound
Jun 14 21:52:50 <[g2]> and listen to the sound of our programs running :)
Jun 14 21:54:52 <rwhitby-treo> I think push reset button on power on to boot cramfs is a good option
Jun 14 21:55:30 <[g2]> rwhitby-treo, is that what openwrt is doing today ?
Jun 14 21:55:45 <rwhitby-treo> yep
Jun 14 21:55:49 <[g2]> I'd probably just go with boot different jffs2
Jun 14 21:56:03 <rwhitby-treo> either
Jun 14 21:56:26 <[g2]> adding the second jffs2 would be pretty trivial
Jun 14 21:56:37 <rwhitby-treo> whatever is in the config.  default to cramfs if no config
Jun 14 21:56:50 <[g2]> APEX doesn't know cramfs yet
Jun 14 21:57:12 <rwhitby-treo> config has two rootfs locations, and default to ro cramfs if no config
Jun 14 21:57:42 <[g2]> rwhitby-treo, I think you're missing the point about the cramfs
Jun 14 21:58:12 <rwhitby-treo> [g2] is so stuck in the present ... :-)
Jun 14 21:58:31 <[g2]> rwhitby-treo, funny how we've switch :)
Jun 14 21:58:55 <rwhitby-treo> he he
Jun 14 21:59:00 <rwhitby-treo> indeed
Jun 14 21:59:44 <[g2]> either that or all the stuff I've been working has taken an extra 9 months :)
Jun 14 22:09:00 <rwhitby-treo> [g2]: apex and openslug are gonna rock
Jun 14 22:10:22 <[g2]> rwhitby-treo, apex and openslug rock :)
Jun 14 22:10:25 <rwhitby-treo> we just need to solve that irq26 problem ...
Jun 14 22:10:39 <jacques> this is gonna sound really weird
Jun 14 22:10:46 <jacques> but oh well, I don't care
Jun 14 22:10:56 <[g2]> actually that'd be nice, but it can be masked by a boot parameter
Jun 14 22:11:04 <rwhitby-treo> [g2]: gonna rock for more people than just you and bw
Jun 14 22:11:37 <jacques> I can't think of any reason for it, but my 2.3.5 openslug seems to do the irq26 think much less than my other one
Jun 14 22:11:48 <[g2]> rwhitby-treo, I'm just teasing you while i've got the opportunity, since it comes by so rarely
Jun 14 22:12:23 <[g2]> next time I boot with APEX i'll try the nodebugirq
Jun 14 22:12:31 <rwhitby-treo> enjoy it while it lasts - I have a free slug with jtag now
Jun 14 22:12:56 <[g2]> the only real problem with Openslug is the reset/reboot issue
Jun 14 22:13:15 <[g2]> the irq 26 is a moderate problem
Jun 14 22:13:25 <[g2]> the reset/reboot is Critical
Jun 14 22:14:04 <[g2]> but I think the reset/reboot problem can be found and fixed more easily
Jun 14 22:14:26 <[g2]> jacques, what's the status of 2.6.12 anyway ?
Jun 14 22:16:23 <jacques> from what I hear there are so many fundamental changes it will be months before it's working
Jun 14 22:16:37 <jacques> especially with rmk being a little )%*(($%$
Jun 14 22:17:49 <jacques> that's mr "I never think about pxa or xscale anymore" rmk to you
Jun 14 22:18:07 <ka6sox> hmmm....
Jun 14 22:18:10 <jacques> (I assume you saw that quote)
Jun 14 22:18:26 <ka6sox> um...I didn't
Jun 14 22:18:44 <jacques> on lak
Jun 14 22:27:32 <rwhitby-treo> is a new arm-linux maintainer called for?
Jun 14 22:28:20 <[g2]> well nite all
Jun 14 22:28:27 <[g2]> rwhitby-treo, have fun with APEX :)
Jun 14 22:28:53 <jacques> rwhitby-treo, you aren't the first (or the 100th) person to ask that question
Jun 14 22:28:53 <rwhitby-treo> night g2
Jun 14 22:29:03 <jacques> this has been going on for years
Jun 14 22:32:34 <jacques> it seems the more ppl who start to understand the arm parts of the kernel, the more pissy rmk gets
Jun 14 22:32:46 <rwhitby-treo> heh
Jun 14 22:33:11 <rwhitby-treo> the child is leaving home ...
Jun 14 22:33:19 <jacques> must be
Jun 14 22:57:23 <rwhitby-treo> back later
Jun 14 22:58:40 <eno> any1 seen this: fetch http://www.oesources.org/source/current/flex-2.5.31.tar.bz2 404 Not Found
Jun 14 22:59:19 <ka6sox> yep...
Jun 14 22:59:24 <eno> i'm following "OpenEmbedded/GettingStarted" for the 1st time
Jun 14 22:59:31 <eno> any work around?
Jun 14 23:00:04 <ka6sox> yeah...download the svn tarball of unslung firmware from either sf.net or berlios.de
Jun 14 23:00:31 <ka6sox> it has the url for a download target with that source in it (it will put it in your downloads directory.)
Jun 14 23:00:31 <eno> k
Jun 14 23:00:55 <ka6sox> either the 4.20 or 5.5 tarball have them.
Jun 14 23:01:13 <ka6sox> (its in the file release area)
Jun 14 23:01:51 <eno> thx, will try that
Jun 14 23:01:54 <ka6sox> there is a make target for the downlaod
Jun 14 23:05:50 <eno> is the key simply get the flex-2.5.31.tar.bz2 ?
Jun 14 23:08:40 <ka6sox> yep
Jun 14 23:09:10 <ka6sox> lemme see where we stashed it.
Jun 14 23:15:23 <eno> will this http://bindist.finkmirrors.net/bindist/dists/fink-0.7.1/main/source/devel/flex-2.5.31.tar.bz2 work?
Jun 14 23:15:56 <ka6sox> anything that has the right source
Jun 14 23:16:19 <ka6sox> just put it in your downloads directory so it can be found.
Jun 14 23:16:47 <eno> seems working, it's downloading the following tarballs now :-)
Jun 14 23:16:59 <ka6sox> :)
Jun 14 23:17:21 <jacques> `night all
Jun 14 23:21:31 <ka6sox> nite jacques 
Jun 14 23:27:03 <eno> just started a 'bitbake unslung-image' build
Jun 14 23:28:58 <eno> nite all
Jun 14 23:29:12 <eno> thanks again ka6sox
Jun 14 23:29:53 <ka6sox> eno-zzz, np...
Jun 14 23:40:13 <mr_claus> morning
Jun 14 23:40:32 <ka6sox> howdy
Jun 14 23:40:43 <mr_claus> can anybody give me access to the mailinglist please, i tried to join but i didn't get the email from yahoo :/
Jun 14 23:40:51 <mr_claus> same problem like yesterday
Jun 14 23:41:16 <ka6sox> dunno why you can't get the email...check your junk mail folder
Jun 14 23:41:50 <mr_claus> i checked it, nothing in there
Jun 14 23:42:18 <ka6sox> I'm in the middle of something right now but when I'm done I can look at it.
Jun 14 23:44:07 <mr_claus> i checked my mailserver now, no incoming mail from yahoo
Jun 14 23:44:18 <ka6sox> k
Jun 14 23:53:56 <mr_claus> i will be not on my computer the next hours, i will tell you if i'm back, it seems to be a yahoo problem because it should be possible to register the email address to the account, yahoo should send an email which should be confirmed, but i never got an email
Jun 14 23:54:05 <mr_claus> i think thats the reason why it's not working
Jun 14 23:55:50 * mr_claus is away: I'm busy
**** ENDING LOGGING AT Tue Jun 14 23:59:56 2005