**** BEGIN LOGGING AT Sat Jun 25 02:59:57 2011
Jun 25 05:48:45 <PaulFertser> lindi-_: please consider this instead: http://paste.debian.net/120917/
Jun 25 06:10:50 <PaulFertser> Or even puts("\n") for the first hunk.
Jun 25 06:16:02 <PaulFertser> lindi-_: ("Starting") there's no difference code-wise for any device. Probably it's your kernel from the second partition with early_printk enabled invalidates the tx fifo before it has a chance of outputting that string.
Jun 25 06:18:02 <PaulFertser> (wild guess)
Jun 25 06:42:22 <josch> lindi-_: JaMa sent me his defconfig - try this kernel image: http://mister-muffin.de/p/iiS9.bin
Jun 25 06:42:31 <josch> (boots here)
Jun 25 06:49:27 <PaulFertser> josch: hey :D
Jun 25 06:49:48 <josch> PaulFertser: hi pavel :)
Jun 25 06:50:12 <josch> long time no see :)
Jun 25 06:50:20 <PaulFertser> josch: how's life?
Jun 25 06:50:41 <josch> PaulFertser: as you can see: finally having some time to give my gta02 some lovin ^^
Jun 25 06:50:55 <josch> also i'm preparing for the arrival of the gta04 early adopter
Jun 25 06:51:18 <PaulFertser> :)) toys ftw
Jun 25 06:51:41 <josch> PaulFertser: recently graduated and only have some small job over the summer - so enough time for toys :)
Jun 25 06:53:31 <PaulFertser> josch: what degree have you got? 
Jun 25 06:53:52 <josch> bachelor of science - master comes now :)
Jun 25 06:54:00 <PaulFertser> Yay
Jun 25 06:54:05 <josch> \o/
Jun 25 06:55:57 <josch> PaulFertser: i hope everything is alright for you as well?
Jun 25 08:48:01 <GNUtoo> hi leviathan__ 
Jun 25 08:48:07 <GNUtoo> I've bad news and good news
Jun 25 08:50:39 <DocScrutinizer> anybody got PM spammed here?
Jun 25 08:51:12 <DocScrutinizer> no?
Jun 25 08:51:16 <DocScrutinizer> good
Jun 25 08:52:26 <DocScrutinizer> btw: /mode <yourusername> +R seems to help
Jun 25 08:52:39 <DocScrutinizer> just in case of PM flooding
Jun 25 09:02:50 <GNUtoo> ping leviathan__ 
Jun 25 09:19:02 <GNUtoo> leviathan__, ping me as soon as you can
Jun 25 12:57:21 <GNUtoo> mickey|daddy, hi
Jun 25 14:10:53 <GNUtoo> hi DocScrutinizer 
Jun 25 14:11:03 <GNUtoo> you said that long time ago:
Jun 25 14:11:06 <DocScrutinizer> yo hi
Jun 25 14:11:18 <GNUtoo> Jun 10 00:05:25 <DocScrutinizer>        when writei() side calls callback function: if (writeptr-readptr)<chunksize_for write_N then exit; else write N bytes from ringbuffer to writei(); update readptr; exit
Jun 25 14:11:24 <GNUtoo> what is readptr and writeptr
Jun 25 14:11:33 <GNUtoo> there are so many read and write pointers
Jun 25 14:11:42 <DocScrutinizer> a local variable
Jun 25 14:11:50 <DocScrutinizer> pointing into your ringbuffer
Jun 25 14:12:07 <GNUtoo> let's say I've a ring buffer with 2*160 frames
Jun 25 14:12:18 <GNUtoo> let's call it buffer[0] and buffer[1]
Jun 25 14:12:33 <GNUtoo> should I do it that way?
Jun 25 14:12:36 <DocScrutinizer> your buffer can be much larger, doesn't matter
Jun 25 14:13:08 <GNUtoo> with sizeof(buffer[0]) beeing 160
Jun 25 14:13:17 <DocScrutinizer> but yes, going for absolute minimum size which is 2*1 chunk still is applicable
Jun 25 14:14:15 <GNUtoo> then what is chunksize?
Jun 25 14:14:31 <GNUtoo> it's not N = readi/writei () ?
Jun 25 14:14:31 <DocScrutinizer> actually your buffer probably should be 2* chunksize + 2frames for separation, minimum
Jun 25 14:15:01 <DocScrutinizer> chunksize is the maximum size you get by one call to readi()
Jun 25 14:15:21 <DocScrutinizer> i.e. alsa buffer size
Jun 25 14:15:25 <GNUtoo> ok
Jun 25 14:15:28 <DocScrutinizer> or cmt-buffer
Jun 25 14:15:29 <GNUtoo> then I'll re-explain
Jun 25 14:15:55 <GNUtoo> can I do something that uses 2 buffers like buffer[0] and buffer[1] ?
Jun 25 14:16:05 <GNUtoo> and doing a not on 0 or 1
Jun 25 14:16:17 <GNUtoo> but then the pointer would be 0 or 1
Jun 25 14:16:22 <GNUtoo> not something like 160
Jun 25 14:16:23 <DocScrutinizer> yes, though I suggest to use one buffer instead of 2
Jun 25 14:16:38 <GNUtoo> ah like pointer beeing 160
Jun 25 14:16:46 <DocScrutinizer> and use frame-pointers instead of buffer pointers
Jun 25 14:17:03 <GNUtoo> so when you update readptr/writeptr you update from chunk size
Jun 25 14:17:15 <GNUtoo> like ptr+= N
Jun 25 14:17:21 <GNUtoo> with n = readi()
Jun 25 14:17:31 <GNUtoo> ok
Jun 25 14:17:34 <GNUtoo> thanks a lot
Jun 25 14:17:35 <DocScrutinizer> you take out 160 frames from ringbuffer and advance readpointer by 160
Jun 25 14:17:58 <DocScrutinizer> you insert 400 frames and advance writeptr by 400
Jun 25 14:18:03 <GNUtoo> ok
Jun 25 14:18:14 <DocScrutinizer> you always take care about wrap-around
Jun 25 14:18:23 <GNUtoo> yes that's in your code
Jun 25 14:18:29 <GNUtoo> if readptr -  writeptr
Jun 25 14:18:31 <DocScrutinizer> and make sure readptr doesn't get in front of writeptr
Jun 25 14:18:36 <GNUtoo> *pseudocode
Jun 25 14:19:07 <DocScrutinizer> nor writepointer runs over readptr
Jun 25 14:19:22 <GNUtoo> ok
Jun 25 14:19:34 <DocScrutinizer> both situations signal a xrun condition that needs handling
Jun 25 14:19:51 <GNUtoo> snd_pcm_recover
Jun 25 14:19:58 <GNUtoo> ?
Jun 25 14:20:01 <DocScrutinizer> usually not
Jun 25 14:20:37 <GNUtoo> then discard frame?
Jun 25 14:20:40 <DocScrutinizer> as you'd just fill send-buffer with silence instead of data when your readptr runs int writeotr
Jun 25 14:20:47 <GNUtoo> ok
Jun 25 14:21:16 <DocScrutinizer> when writeprtr runs into readptr, you probably want to push readptr ahead accordingly
Jun 25 14:21:37 <DocScrutinizer> measn you drop most oldest samples
Jun 25 14:21:43 <DocScrutinizer> means*
Jun 25 14:22:00 <DocScrutinizer> don't forget to set a mutex when doing that
Jun 25 14:22:05 <GNUtoo> yes
Jun 25 14:22:12 <GNUtoo> I was thinking of mutexes
Jun 25 14:22:24 <GNUtoo> not here precisely tough
Jun 25 14:22:30 <DocScrutinizer> only when writing side messes with readptr
Jun 25 14:22:48 <GNUtoo> ah ok
Jun 25 14:22:51 <DocScrutinizer> you need to secure this messing with readptr by a mutex
Jun 25 14:23:00 <GNUtoo> ok
Jun 25 14:23:36 <DocScrutinizer> and, as mentioned, I suggest to make size of ringbuffer at least 3 chunks, not 2
Jun 25 14:23:49 <GNUtoo> ok
Jun 25 14:24:16 <DocScrutinizer> this way you get a bit of jitterbuffer
Jun 25 14:24:44 <DocScrutinizer> you can go up to aritrary length with buffer size, if you want more jiterbuffer
Jun 25 14:24:53 <DocScrutinizer> of course will increase latency
Jun 25 14:25:54 <DocScrutinizer> but I heard even latency of internet-via-sat (2 * 36000km speed of light one way) isn't too much for e.g VoIP
Jun 25 14:25:56 <DocScrutinizer> ;-)
Jun 25 14:26:19 <DocScrutinizer> so don't worry about 50 or 100ms latency
Jun 25 14:26:27 <DocScrutinizer> it makes your job more easy
Jun 25 14:26:58 <DocScrutinizer> about to go afk... so anything open?
Jun 25 14:27:09 <GNUtoo> ok
Jun 25 14:27:13 <DocScrutinizer> :-)
Jun 25 14:27:28 <GNUtoo> I'll implement what you said as pseudocode
Jun 25 14:27:34 <GNUtoo> and then go to real code
Jun 25 14:27:48 <GNUtoo> maybe I'll show you the pseudocode when it's done
Jun 25 14:27:53 <DocScrutinizer> I'll joyfully review it
Jun 25 14:27:59 <GNUtoo> thanks a lot
Jun 25 14:28:03 <DocScrutinizer> yw
Jun 25 14:28:12 <DocScrutinizer> many thanks for taking care
Jun 25 14:37:27 <DocScrutinizer> no shellscript hackers around?
Jun 25 14:38:45 <apt> not me, doc
Jun 25 14:44:00 <lindi-_> no, today is a python day :)
Jun 25 14:44:36 <DocScrutinizer> question: is there a better shellscript way i.e a simpler command to patch binary files via script, than #> od $params infile.bin | sed $params | printf $format >patchedfile.bin  ?
Jun 25 14:46:05 <DocScrutinizer> I'm perfectly fine with answers like "none I know of" from hackers that think they know a bit about shellscripting and patching
Jun 25 14:47:35 <DocScrutinizer> lindi-_: probably it'd be friggin simple to write a py script to patch bytes at random addr in a binary file
Jun 25 14:48:49 <DocScrutinizer> though I'd accept an answer along #> python -c "bla bla bla"; though I had to reject it in the end as I can't use python for my case
Jun 25 14:49:29 <DocScrutinizer> s/though/so/
Jun 25 14:49:29 <apt> DocScrutinizer meant: so I'd accept an answer along #> python -c "bla bla bla"; though I had to reject it in the end as I can't use python for my case
Jun 25 14:50:25 <DocScrutinizer> it's quite baffling there seems to be no binary analog to patch cmd
Jun 25 14:51:10 <DocScrutinizer> so I was wondering if maybe I missed sth obvious and you actually could use patch for binaries as well
Jun 25 14:51:19 <lindi-_> DocScrutinizer: xdelta3
Jun 25 14:51:30 <DocScrutinizer> o.O
Jun 25 14:51:46 <lindi-_> and python-xdelta3 too
Jun 25 14:51:51 <DocScrutinizer> sounds intriguing, though I bet it's not unix toolbox :-)
Jun 25 14:51:59 <lindi-_> I use the python API. there's a command-line tool for old *nix folks too
Jun 25 14:52:09 <lindi-_> DocScrutinizer: the problem with unix is that you can't improve it.
Jun 25 14:52:27 <DocScrutinizer> why not?
Jun 25 14:52:48 <lindi-_> DocScrutinizer: how would you improve it?
Jun 25 14:53:04 <DocScrutinizer> add cmds to unix toolbox
Jun 25 14:53:04 <lindi-_> DocScrutinizer: you write a new tool and people refuse to use it since it is not in unix :)
Jun 25 14:53:11 <lindi-_> DocScrutinizer: how do you do that?
Jun 25 14:53:12 <DocScrutinizer> yeah
Jun 25 14:53:20 <DocScrutinizer> you're right
Jun 25 14:53:57 <DocScrutinizer> that's why I thought maybe somebody augmented patch by some binary mode
Jun 25 14:54:11 <lindi-_> that'd be evil GNU extension then :
Jun 25 14:54:21 <lindi-_> and BSD people would hate you :P
Jun 25 14:54:23 <DocScrutinizer> so it could use cmp input instead of diff input
Jun 25 14:55:03 <DocScrutinizer> darn, binutils should have a tool since 1956 :-S
Jun 25 14:55:17 <lindi-_> DocScrutinizer: cmd is quite bad if you insert a byte in the middle
Jun 25 14:55:26 <lindi-_> cmp
Jun 25 14:55:28 <DocScrutinizer> not *everything* to patch is a textfile
Jun 25 14:55:53 <DocScrutinizer> nonono, no insertion/deletion of bytes
Jun 25 14:56:08 <DocScrutinizer> simple replacing
Jun 25 14:56:12 <DocScrutinizer> nothing else
Jun 25 14:56:35 <angelox|laptop> GNUtoo; Could fix my problems with kernel power,i was installing from extras-devel,i installed from stable and worked..at least without kexecboot,let me try with it
Jun 25 14:56:35 <lindi-_> anyways, just use xdelta3. it's quite good
Jun 25 14:57:04 <DocScrutinizer> though that patch --binary-mode of course could learn a extended syntax to also insert/delete/modify bytes
Jun 25 14:57:42 <DocScrutinizer> angelox|laptop: ECHAN?
Jun 25 14:58:22 <DocScrutinizer> or did I miss sth recently? ;-D
Jun 25 14:58:22 <GNUtoo> angelox|laptop, ok, btw I'm currently busy inproving sound during calls
Jun 25 14:58:41 <GNUtoo> angelox|laptop, note that you need to use a scenario to make it work right now
Jun 25 14:58:54 <DocScrutinizer> kernel power, extras-devel, sounds all rather maemish
Jun 25 14:58:55 <GNUtoo> (just in case you wanted to try)
Jun 25 14:59:10 <GNUtoo> DocScrutinizer, he has an n900 and wants dual-boot
Jun 25 14:59:15 <DocScrutinizer> aaah
Jun 25 14:59:38 <GNUtoo> so I worked on dual-boot in kexecboot for him
Jun 25 14:59:48 <angelox|laptop> btw GNUtoo: i think i know why uboot didn't work,every try when i was with it my cover was opened :\
Jun 25 14:59:56 <GNUtoo> but....there is 1 minute delay between the kernel booting and kexec -e
Jun 25 15:00:04 <GNUtoo> *kexec -e and the kernel booting
Jun 25 15:00:17 <GNUtoo> angelox|laptop, ok
Jun 25 15:01:03 <DocScrutinizer> lindi-_: so do you know of any prev instance where somebody patched binaries via a script, and how he did it?
Jun 25 15:01:27 <DocScrutinizer> NB a script to deploy
Jun 25 15:03:42 <lindi-_> nothing that I could recommend :)
Jun 25 15:11:32 <DocScrutinizer> I'm doing it with od, sed, and printf :-S
Jun 25 15:11:40 <DocScrutinizer> pff
Jun 25 15:11:50 <DocScrutinizer> ~lart unix
Jun 25 15:11:50 * apt beats the living hamstercrap out of unix
Jun 25 15:26:48 <GNUtoo> http://pastie.org/2120978
Jun 25 15:27:00 <GNUtoo> I wonder if the design could be simplified
Jun 25 15:27:14 <GNUtoo> since the modem buffers always seem to be 160
Jun 25 15:27:47 <GNUtoo> I think that even the pukeaudio plugin hardcode 160 if I remember well
Jun 25 15:27:52 <GNUtoo> *pulseaudio
Jun 25 15:28:44 <GNUtoo> ah sorry I forgott the mutexes
Jun 25 15:29:14 <GNUtoo> so it should be "mutex and so push readptr ahread because of collision "
Jun 25 15:47:11 <GNUtoo> leviathan__, ping
Jun 25 15:51:20 <GNUtoo> http://pastie.org/2121043 is better 
Jun 25 16:07:25 <GNUtoo> mickey|daddy, ping
Jun 25 16:07:31 <GNUtoo> why is that commented in the alsa vapi:
Jun 25 16:07:37 <GNUtoo>         //public int snd_async_add_pcm_handler( snd_async_handler_t **handler, snd_pcm_t *pcm, snd_async_callback_t callback, void *private_data );
Jun 25 16:07:42 <GNUtoo>         //public PcmDevice async_handler_get_pcm( snd_async_handler_t *handler );
Jun 25 16:07:47 <GNUtoo> I need theses functions
Jun 25 16:10:15 <GNUtoo> I'll extern them in the mean time
Jun 25 16:13:11 <GNUtoo> sigh
Jun 25 16:14:37 <GNUtoo> DocScrutinizer, can I use a blocking thread instead?
Jun 25 16:15:00 <GNUtoo> because I fear that binding the snd_async function will be hard
Jun 25 16:16:18 <DocScrutinizer> GNUtoo: sure
Jun 25 16:16:25 <GNUtoo> ok
Jun 25 16:17:16 <DocScrutinizer> while true; do readi_blocking(); write-to-buffer; done
Jun 25 16:21:27 <PaulFertser> josch: yes, quite so :)
Jun 25 16:23:34 <PaulFertser> lindi-_: (qi patch) please consider this instead: http://paste.debian.net/120917/ . Or even puts("\n") for the first hunk.
Jun 25 16:23:53 <PaulFertser> lindi-_: ("Starting") there's no difference code-wise for any device. Probably it's your kernel from the second partition with early_printk enabled invalidates the tx fifo before it has a chance of outputting that string. (wild guess)
Jun 25 16:33:00 <DocScrutinizer> great! :-/  xargs printf is about 10,000 times faster than xarg -Ixxx printf
Jun 25 16:34:14 * angelox|laptop thinks that OE don't want to build anything today. "ERROR: Unable to determine endianness for architecture 'INVALID'"
Jun 25 16:40:20 <DocScrutinizer> lindi-_: PaulFertser: http://paste.debian.net/120972/
Jun 25 16:40:23 <DocScrutinizer> :-o
Jun 25 16:43:19 <DocScrutinizer> suggestions welcome how to convert string 0xFF to a byte in a file
Jun 25 16:46:45 <josch> DocScrutinizer: use xxd
Jun 25 16:50:23 <DocScrutinizer> josch: not in busybox
Jun 25 16:50:32 <DocScrutinizer> but thanks :-)
Jun 25 16:58:00 <PaulFertser> DocScrutinizer: what about printf '\xFF' ?
Jun 25 17:00:56 <DocScrutinizer> trying to get it done with sed
Jun 25 17:01:24 <DocScrutinizer> xarg and all that makes it a bit awkward
Jun 25 17:04:52 <PaulFertser> DocScrutinizer: yes, simple sed should be enough to s/0x/\\x/g
Jun 25 17:11:47 <DocScrutinizer> GRRRRRRR
Jun 25 17:14:28 <DocScrutinizer> the pipe strips the " from "\x77", xargs converts something or dunno, output is NULL almost
Jun 25 17:17:45 <PaulFertser> DocScrutinizer: you seem to be trying to invent a shell script to do binary patching, i see :)
Jun 25 17:20:53 <PaulFertser> DocScrutinizer: you can see here how the guy's using perl to do binary patching in an actual project (it does actually support several different versions of the binary because it doesn't pay attention to addresses, only to the contents): http://nawcom.com/rc.cdrom
Jun 25 17:22:23 <DocScrutinizer> I 'just' need a streaming filter "\0xNN"-><binary char>
Jun 25 17:22:41 <DocScrutinizer> drives me nuts
Jun 25 17:22:55 <DocScrutinizer> od -Ax -tx1 -w1 -v history.sh | sed "s/000001 3d/000001 99/; s/000003 28/000003 99/" | cut -d " " -f2 | sed 's/\(.*\)/"\\x\1\\n"/' | xargs printf %s
Jun 25 17:24:18 <DocScrutinizer> all fine except od giving a bogus last addr line without byte value when hitting EOF (minor problem), and that FRIGGIN "xargs printf"
Jun 25 17:25:20 <DocScrutinizer> od -Ax -tx1 -w1 -v history.sh | sed "s/000001 3d/000001 99/; s/000003 28/000003 99/" | cut -d " (.*\)/"\\x\1"/' | xargs printf %s
Jun 25 17:25:41 <DocScrutinizer> gives me \x6f\x5c\x25\x30\x33\x6f\x5c\x25\x30\x33\x6f.......
Jun 25 17:25:57 <DocScrutinizer> remove the %s and I get nuttin
Jun 25 17:26:44 <DocScrutinizer> I guess busybox doesn't support perl
Jun 25 17:26:58 <DocScrutinizer> I need that for an install script on maemo
Jun 25 17:27:10 <PaulFertser> DocScrutinizer:  xargs -n 1 printf
Jun 25 17:27:18 <DocScrutinizer> o.O
Jun 25 17:27:46 <DocScrutinizer> WOOOHOOOOO
Jun 25 17:27:49 <DocScrutinizer> thanks
Jun 25 17:27:53 <PaulFertser> :)
Jun 25 17:28:00 <DocScrutinizer> running to buy some milk - ttyl
Jun 25 17:28:05 <PaulFertser> See youu.
Jun 25 17:32:43 <PaulFertser> DocScrutinizer: you do not need \n in the last sed i guess and skip cut and the second sed:
Jun 25 17:32:47 <PaulFertser> od -Ax -tx1 -w1 -v .bash_profile | sed 's/000001 3d/000001 99/; s/000003 28/000003 99/; s/.* \(.*\)$/"\\x\1"/' | xargs -n 1 printf
Jun 25 17:33:33 <DocScrutinizer> yeah cruft
Jun 25 17:37:31 <PaulFertser> I'd probably try using single awk instance instead of running printf for every single byte.
Jun 25 17:54:28 <DocScrutinizer> :-x
Jun 25 17:54:47 * DocScrutinizer headdesks
Jun 25 18:04:20 <PaulFertser> I'm near getting it to work
Jun 25 18:11:04 <DocScrutinizer> :-D
Jun 25 18:11:58 <DocScrutinizer> I'm having fun with man awk, something as useless as trying to stop breathing - the effects vanish after short time
Jun 25 18:12:02 <DocScrutinizer> either way
Jun 25 18:14:40 <PaulFertser> DocScrutinizer: od -Ax -tx1 -w1 -v .bash_profile | awk '/000001 20/ { $0 = "000001 32"} /000003 62/ { $0 = "000003 33"} { gsub(/^.* /, "0x"); printf "%c", strtonum($0) }
Jun 25 18:14:54 <DocScrutinizer> hehe
Jun 25 18:15:20 <DocScrutinizer> now let's digest this flavour if gibberish ;-)
Jun 25 18:15:41 <DocScrutinizer> gsub, hmmm
Jun 25 18:19:48 <DocScrutinizer> damn cute
Jun 25 18:20:38 <DocScrutinizer> actually awesome
Jun 25 18:21:04 <DocScrutinizer> now I'll extend it to catch od's fallout at end of file
Jun 25 18:21:24 <GNUtoo> hi mickey|office 
Jun 25 18:21:28 <GNUtoo> how to do that in vala:
Jun 25 18:21:38 <GNUtoo> pcmin.writei(from_modem_to_writei + writeiWriteptr ,FCOUNT);
Jun 25 18:21:49 <GNUtoo> that is to + a pointer
Jun 25 18:22:13 <GNUtoo> ping leviathan__ 
Jun 25 18:22:39 <GNUtoo> maybe I should leave mickey|daddy alone and ask in #vala
Jun 25 18:23:30 <DocScrutinizer> GNUtoo: are you sure this will pan out?
Jun 25 18:23:43 <GNUtoo> what does pan out means?
Jun 25 18:23:57 <GNUtoo> I wrote the code and it's quite complex now
Jun 25 18:24:04 <DocScrutinizer> dunno if you can hand over a part of a ringbuffer to ... err why not. hmmm
Jun 25 18:24:06 <GNUtoo> and different from the pseudocode
Jun 25 18:25:13 <DocScrutinizer> I still wonder why you're doing it in vala
Jun 25 18:25:17 <mickey|daddy> GNUtoo: if all else fails, cast cast cast :)
Jun 25 18:25:19 <mickey|daddy> see here:
Jun 25 18:25:36 <GNUtoo> cast to void*?
Jun 25 18:25:37 <mickey|daddy> http://pastie.org/2121563
Jun 25 18:25:56 <GNUtoo> ok thanks
Jun 25 18:26:13 <mickey|daddy> surely looks ugly, but if it works, just do it :)
Jun 25 18:26:14 <DocScrutinizer> I for one probably would prefer assembler to vala ;-)
Jun 25 18:26:51 <GNUtoo> lol
Jun 25 18:28:31 <PaulFertser> tough guy
Jun 25 18:28:32 <PaulFertser> :)
Jun 25 19:08:01 <DocScrutinizer> PaulFertser: that awk script is fast as hell ;-)
Jun 25 19:08:13 <DocScrutinizer> nice
Jun 25 19:08:35 <GNUtoo> leviathan__, ping
Jun 25 19:08:37 <DocScrutinizer> alas now I lost idea which file to patch by using it :-/
Jun 25 19:15:05 <angelo|compiling> GNUtoo: kexecboot can boot a fiasco-based kernel,or i need extract it and turn into a zImage?
Jun 25 19:15:26 <GNUtoo> it can handle only a zImage or uImage
Jun 25 19:15:30 <GNUtoo> and only power kernel
Jun 25 19:16:00 <angelo|compiling> ok,because i extracted a kernel from the deb package from kernel-power and that there's a fiasco kernel..
Jun 25 19:35:04 <angelo|compiling> GNUtoo: Can i have two entries in a single boot.cfg for two kernels testing?
Jun 25 19:35:40 <GNUtoo|htcdream> yes
Jun 25 19:35:55 <GNUtoo|htcdream> look at kexecboot website for more infos
Jun 25 19:36:42 <angelo|compiling> ok,thank you,and sorry if i ping'ed you in a wrong time
Jun 25 19:37:02 <GNUtoo|htcdream> sh?
Jun 25 19:37:08 <GNUtoo|htcdream> np
Jun 25 19:37:16 <GNUtoo|htcdream> s/sh//
Jun 25 19:59:11 * angelo|compiling gives up
Jun 25 20:12:23 <GNUtoo|htcdream> ?
Jun 25 20:14:55 <angelo|compiling> camera,bluetooth,lock button and kernel-power..nothing else kernel-power works
Jun 25 20:19:29 <GNUtoo> complain to #maemo
Jun 25 20:20:04 <angelo|compiling> already did that
Jun 25 20:20:13 <GNUtoo> what did they say?
Jun 25 20:20:16 <GNUtoo> it's not normal
Jun 25 20:20:22 <GNUtoo> for me everything works
Jun 25 20:20:48 <GNUtoo> under maemo
Jun 25 20:21:06 <GNUtoo> I can abandon the alsa plugin for a minute and verify that if you want
Jun 25 20:21:20 <angelo|compiling> no,that's not needed..
Jun 25 20:21:32 <angelo|compiling> my english is too bad to explain problems i guess...
Jun 25 20:22:13 <GNUtoo> what did they say?
Jun 25 20:22:18 <angelo|compiling> they said to me re-install nokia stock kernel and install kernel-power from wiki..i did that and it worked,but without kexecboot...
Jun 25 20:22:54 <GNUtoo> and all that works with stock kernel?
Jun 25 20:23:01 <GNUtoo> does wifi works with kernel power?
Jun 25 20:25:32 <angelo|compiling> yes
Jun 25 20:25:59 <angelo|compiling> with kernel power + kexecboot wifi works
Jun 25 20:26:41 <GNUtoo> hmmm
Jun 25 20:26:57 <GNUtoo> are you able to investigate more?
Jun 25 20:27:20 <angelo|compiling> yes
Jun 25 20:27:42 <GNUtoo> because look at /dev/input/
Jun 25 20:27:44 <GNUtoo> with evtest
Jun 25 20:27:51 <GNUtoo> with kenrel-power booted
Jun 25 20:27:55 <GNUtoo> ah no
Jun 25 20:28:01 <GNUtoo> they use another system
Jun 25 20:28:03 <GNUtoo> I remember
Jun 25 20:28:07 <GNUtoo> with gpio_switch
Jun 25 20:28:10 <GNUtoo> or something like that
Jun 25 20:28:58 <angelo|compiling> i guess is something related about modules...
Jun 25 20:29:02 <angelo|compiling> not sure
Jun 25 20:29:04 <GNUtoo> yes
Jun 25 20:29:28 <GNUtoo> first do a dmesg
Jun 25 20:29:36 <GNUtoo> to look if everything is fine
Jun 25 20:30:10 <angelo|compiling> ok
Jun 25 20:33:22 <angelo|compiling> seems like ok...
Jun 25 20:35:12 <angelo|compiling> for some reason camera worked now 
Jun 25 20:35:20 <GNUtoo> http://git.freesmartphone.org/?p=cornucopia.git;a=blob_plain;f=fsodeviced/src/plugins/gpio_input/plugin.vala;hb=HEAD
Jun 25 20:35:30 <GNUtoo> that's the system used by the 2.6.28
Jun 25 20:35:57 <GNUtoo> let me look in the nokia900 config
Jun 25 20:36:48 <angelo|compiling> hmm
Jun 25 20:36:50 <GNUtoo> basically there is slide,kb_lock,cam_focus,cam_launch,cam_shutter and headphone
Jun 25 20:37:06 <GNUtoo> don't be afraid by the fact that I'm pointing at fso code
Jun 25 20:37:07 <angelo|compiling> /dev/input/pwrbutton answer me when i press it
Jun 25 20:37:11 <GNUtoo> ok
Jun 25 20:37:25 <GNUtoo> power buttton is the normal input system
Jun 25 20:37:29 <angelo|compiling> but there's no *lock* event...
Jun 25 20:37:36 <GNUtoo> that's normal
Jun 25 20:37:40 <GNUtoo> I can re-explain
Jun 25 20:37:53 <GNUtoo> some buttons uses a non-standard gpio-input thing
Jun 25 20:37:58 <GNUtoo> under maemo
Jun 25 20:38:04 <GNUtoo> and with 2.6.28
Jun 25 20:38:11 <GNUtoo> we used the 2.6.28 at the beginning
Jun 25 20:38:17 <GNUtoo> now we use 2.6.37
Jun 25 20:38:30 <angelo|compiling> sorry didn't read you saying about another way that maemo uses to read things*
Jun 25 20:38:41 <GNUtoo> yes
Jun 25 20:38:44 <GNUtoo> that's gpio_input
Jun 25 20:40:43 <angelo|compiling> GNUtoo: strange: when i press lock button nothing happens,but when i press it and do a "dmesg | tail" the screen locks normally...
Jun 25 20:40:50 <GNUtoo> robot voice now
Jun 25 20:41:02 <GNUtoo> strange strange
Jun 25 20:41:16 <angelo|compiling> yes,and dmesg returns like everything is fine "[  702.472106] kb_lock (GPIO 113) is now open/closed" 
Jun 25 20:43:35 <angelo|compiling> if i press lock button and release nothing happen,if i press it and wait about 3 seconds,it locks the screen...
Jun 25 20:48:07 <angelo|compiling> hmmmmm
Jun 25 20:48:17 <angelo|compiling> camera button doesn't work :/
Jun 25 20:48:45 <angelo|compiling> camera closing/opening case doesn't work
Jun 25 20:49:18 <angelo|compiling> s/case/shutter/
Jun 25 20:49:19 <apt> angelo|compiling meant: camera closing/opening shutter doesn't work
Jun 25 20:50:34 <GNUtoo> ah
Jun 25 20:50:39 <angelo|compiling> dmesg | tail makes everything work,camera shutter,camera button and lock button
Jun 25 20:50:51 <GNUtoo> hmmm
Jun 25 20:51:03 <GNUtoo> maybe maemo looks in dmesg for the event
Jun 25 20:51:11 <angelo|compiling> i guess the O.S. isn't accessing correctly gpio nodes...
Jun 25 20:51:12 <GNUtoo> anyway that's an important information
Jun 25 20:52:57 <angelo|compiling> maybe a bad-compilation of kernel-power that makes it don't work nicely,no idea..
Jun 25 20:53:07 <angelo|compiling> i'd like to try a different version of that
Jun 25 20:57:35 <angelo|compiling> GNUtoo: wiki says to me that reflashing a complete firmware image is the best way to upgrade to PR 1.3,and not just flashing kernel,do you thing that would be cool flash everything ?
Jun 25 21:04:43 <GNUtoo> yes, you could try
Jun 25 21:04:47 <GNUtoo> but 
Jun 25 21:04:49 <GNUtoo> beware
Jun 25 21:05:03 <GNUtoo> if something fails
Jun 25 21:05:07 <GNUtoo> you'll need flasher
Jun 25 21:05:14 <GNUtoo> which doesn't work for you
Jun 25 21:05:52 <angelo|compiling> why?
Jun 25 21:06:06 <GNUtoo> because you are flashing the bootloader etc...
Jun 25 21:06:18 <angelo|compiling> i mean,why flasher doesn't work?
Jun 25 21:06:23 <GNUtoo> no idea
Jun 25 21:06:29 <GNUtoo> I don't remember the error
Jun 25 21:06:39 <GNUtoo> you told me it didn't work for you
Jun 25 21:06:50 <angelo|compiling> ah,but now i'm in another distribution,it should work now,i'll flash from it to don't get any error
Jun 25 21:06:52 <GNUtoo> and since it's proprietary you cannot fix it
Jun 25 21:07:09 <GNUtoo> yes but beware because it's proprietary
Jun 25 21:07:34 <angelo|compiling> ok
Jun 25 21:07:38 <angelo|compiling> thank you
Jun 25 21:07:39 <GNUtoo> and I guess it requires root permissions
Jun 25 21:07:54 <GNUtoo> so a program that you don't know what it does with root permissions....
Jun 25 21:08:00 <GNUtoo> that could do virtually anything
Jun 25 21:08:28 <GNUtoo> leviathan__, ping
Jun 25 21:08:41 <GNUtoo> maybe leviathan__ is hidding?
Jun 25 21:09:02 <GNUtoo> I need to talk to you to decide between 2 things:
Jun 25 21:09:07 <GNUtoo> or:
Jun 25 21:09:15 <GNUtoo> I buy a new battery from geeksphone
Jun 25 21:09:19 <GNUtoo> or:
Jun 25 21:09:29 <GNUtoo> you find where you put the battery and re-send it to me
Jun 25 21:09:49 <GNUtoo> I need to decide quickely
Jun 25 21:09:53 <GNUtoo> the battery are cheap
Jun 25 21:09:58 <GNUtoo> and they're in europe
Jun 25 21:18:56 <angelo|compiling> flashed everything...now i'll fight again
**** ENDING LOGGING AT Sun Jun 26 02:59:57 2011