<monofuel>
well i can generate the world on my desktop
<unclouded>
of course the best thing would be to ask them if they'd be willing to target the NN
<monofuel>
it's just got a single programmer who's already buisy enough bugfixing the game
<monofuel>
does the nn have libpng12?
<xdpirate>
alrighty im out for now, thanks for the help unclouded
<unclouded>
it seems you need 10 times the CPU power to emulate another machine, so a 330 MHz NN might be equivalent to a 33 MHz 486
<monofuel>
not sure if it's needed for a bash though
<monofuel>
yea that's what i was mostly worried of
<monofuel>
i figured it was a longshot. my other choice is running it over ssh with wifi
<unclouded>
presumably that will work if the new version can run in terminal mode.  best to check first though!
<monofuel>
best to check first?
<unclouded>
looks like libpng has been ported to OpenWRT, not sure of the version though.  of course if you install Gentoo on SD then you have access to every single Gentoo port
<monofuel>
ooohhhhhh :)
<unclouded>
which would of course include angband, nethack and so on.
<monofuel>
would the next device to follow the nn be a ways off still?
<monofuel>
i just wouldn't want to buy a nn and have something new come out next week :)
<monofuel>
snes-nethack-sdl- what other cool uses are there?
<unclouded>
no idea sorry.  The Ya NanoNote definitely won't be out next week.  if you're considering a NN, you should be aware of the shortcomings such as no USB HOST but don't forget the good point too: it's cheap, it's got a real keyboard that is surprisingly nice to type on, it's incredibly small and it's actually available now, unlike the Pandora
<monofuel>
ah k thx!
<unclouded>
I guess different people will find different uses.  I only want to take a single device with me and that's my phone, so the NN won't be going everywhere with me.  I do think it would be great to take in to the garden with astronomy software and to emulate the C64
<unclouded>
music player is a big one: a music player that actually has a keyboard with which to cue up your next track
<monofuel>
yea
<monofuel>
would the cpu be fast enough for videos?
<unclouded>
not sure.  I think people have done MPEG1 transcoded to 320x240 although I don't know if it would handle MPEG4
<monofuel>
ah- what about mouse controls? does it have a nub/joystick thing like some laptops have or?
<unclouded>
no dedicated mouse but I think people have emulated a mouse with the arrow keys, which are laid out nicely
<monofuel>
does it have java support?
<monofuel>
sorry if most of my questions seem like i'm asking too much from a simple device
<monofuel>
i'm just wondering how far i can go with it
<unclouded>
doesn't look like it's ported to OpenWRT.  not sure if OpenJDK has been ported to MIPS or not
<unclouded>
questions are fine ;)
<monofuel>
how would i go about using a usb:eth cable with it?
<monofuel>
i found a small portable wifi router that i could use
<unclouded>
it comes with a standard USB A to micro B cable.  when you use one to connect your workstation to the NN, the workstation gains a usb0 interface.  The NN already had one.  then you configured those two interfaces and you have a small private network
<unclouded>
then you can set up NAT on your workstation so your NN can get to the Internet via your workstation
<monofuel>
i'm just wondering about connecting it to a wireless router running as a client over eth
<unclouded>
your wireless router would have to support USB HOST
<monofuel>
powered by a small duracell usb charter
<monofuel>
and plug the nn into the wifi point over ethernet
<unclouded>
not sure.  it doesn't say whether it supports USB HOST.  I'm a bit out of my depth here but I'd suggest looking at OpenWRT supported devices and see which have USB HOST
<monofuel>
well, it would connect to the nanonote over ethernet
<monofuel>
is usb host needed for a usb-ethernet conversion?
<monofuel>
ah oh well- too much to carry around anyway
<unclouded>
if you really have to get the NN on to the Internet when you're out and about then SDIO wifi seems the way to go.  if don't mind putting some work in you could look into tethering your phone to the NN's serial port and using GPRS or 3G to access the Internet
<monofuel>
serial port?
<unclouded>
yes, the NN has a serial port.  It would need some electronics to get it tethered to a phone though so not really an option unless you're into that sort of thing
<monofuel>
i don't even have a nice phone to tether to
<monofuel>
i came acrost a couple pages on nn wifi...
<unclouded>
surprisingly many phones can be tethered.  check the gnokii package in your gentoo for more information
<monofuel>
my phone doesn't even have 3g. or text messaging. or anything other than calling.
<unclouded>
which phone is it?
<monofuel>
a really really old nextel one
<unclouded>
maybe you are out of luck there ;)
<monofuel>
i'd like to get an android phone such as the droid, but at minimum wage......
<monofuel>
not sure what other uses i could have for the nn...
<monofuel>
would be cool for nethack and custom-made sdl games on the go :)
<unclouded>
the limit is your imagination ;)Â Â python is installed as standard and pygame is an "opkg install pygame" away if you want to be a road warrior
<monofuel>
ah- forgot all about python! i'm a little more familiar with pygame than perl-sdl actually
<monofuel>
and what about using directFB instead of an x server?
<monofuel>
is that tricky to setup?
<unclouded>
there's nothing to do.  it just goes.  just install pygame and start your python app.  you might need to update to the latest firmware to have python installed as standard.  I got great help from this channel when updating the firmware on my NN on the first day I got it
<monofuel>
ah, so directFB is just automatic for graphical apps?
<monofuel>
sounds suspiciously simple :)
<unclouded>
it really is that simple.  pygame is bound to SDL and SDL uses directfb and directfb uses the kernel frame buffer device
<monofuel>
neat
<unclouded>
so there's another use case: go through all the pygame games and see which run well on the NN ;)
<unclouded>
or write your own, on the device itself if you like
<wolfspraul>
wpwrak: I think your --plot extension to eeschema has not yet gone upstream
<wolfspraul>
in terms of finding the root .sch, I think the .pro file might have a link to it?
<wolfspraul>
at least that would be an official 'pointer' instead of inventing some new way to identify the root. of course we would then need to look for .pro files and follow the link to the root .sch
<Necrosporus>
I have decided to avoid buying Ben NanoNote, because it doesn't have internal bluetooth and is rather slow (only 32M ram, while modern Mandriva works slowly even on 512M RAM computer)
<Necrosporus>
Is it planned to make next generation device with some wireless connectivity, bluetooth or wi-fi?
<Necrosporus>
bt for me is preferable
<wolfspraul>
Necrosporus: there's a lot of discussions around that. personally I think both bt and wi-fi are unlikely.
<Necrosporus>
Than internal USB
<wolfspraul>
it's just very hard to make any progress on the freedom side with these technologies, and if the hardware we are making is just about licensing proprietary technology, we could stop the project and license something really good right away :-)
<Necrosporus>
wolfspraul, I already have some USB dongle which work with gNewSense
<Necrosporus>
I wonder if it's possible to make a usb flash which can be drown into USB port completely
<zear>
you have to get it out somehow
<Necrosporus>
It would be impossible to extract it without special tools, but it would be invisible
<Necrosporus>
If someone would not try to insert another USB device into that port
<zear>
i have also seen wifi dongles that are just slightly bigger from that size bt dongles
<Necrosporus>
Anyway, I want a device for reading books from fb2 files
<Necrosporus>
It should be able to receive files from bluetooth, or at least support some sort of memory cards, it should have some backlight and  good-for-eyes screen
<Necrosporus>
I'm not sure if nano-note screen is good for eyes
<wpwrak>
wolfspraul: (--plot) i know. dick doens't like it because it's not "c++-ish" enough. but he doesn't have anything better either. so it'a just a patch to add.
<zear>
Necrosporus, 320x240 is not a good res for ebooks
<Necrosporus>
I read books from such a screen, it was good enough
<wpwrak>
wolfspraul: (root .sch) the .pro name would be a strong hint. it's difficult to use different names for .pro and .sch
<zear>
additionally nanonte's led placement on the screen is not good for small text reading
<Necrosporus>
Only there is a problem, my PDA is almost dead
<wpwrak>
wolfspraul: btw, i got my ursp2 out of customs yesterday :) didn't get to play with it yet, though
<wpwrak>
Necrosporus: for wirless, it all depends on what chips we can find. if you can find something great for BT (or even wifi), that would help. there are the following issues o consider:
<wolfspraul>
wpwrak: [usrp] great! I saw the wpan commits, very curious what's going to happen next... :-)
<wolfspraul>
about the eeschema, well, we just need to settle on something
<wolfspraul>
we can say we assume thers is exactly 1 .pro file, and it has a RootSch= setting, and that will point to the schematics are are going to plot
<wpwrak>
Necrosporus: 1) can we actually get it. some chips are only sold to large customers, who buy hundreds of thousand. there may also be an up-front fee to pay to license the use of the technology, that can be in the order of 1 million USD, perhaps more. for some wireless technologies.
<wolfspraul>
for plotting I like the --plot, although there are many options in the dialogs, svg, etc. so maybe a more general way to script the UI would be good
<wolfspraul>
but in the end I don't want to talk, I want to make something work on the server now, because we have active projects and automatically generated schematics could help quite a bit
<wolfspraul>
finally - how do we build the kicad binary for the server? if upstream is slow in accepting stuff, and maybe later we want fped or other extensions from you as well, maybe we just build from your kicad sources then?
<wolfspraul>
what is the 'official' way to build "werner's latest kicad"?
<Necrosporus>
wpwrak, there is a simple solution: internal usb host
<wpwrak>
Necrosporus: 2) is there publicly available documentation. we don't have that for everything (e.g., the CPU itself has a semi-closed data sheet), but it makes development a lot easier to have a data sheet we can actually share. if it's in the public, that's also about the only way to make sure you can still maintain your platform after the next company merger or policy change.
<wolfspraul>
Necrosporus: internal USB host is unsatisfactory, at least for me, for a number of reasons:
<wpwrak>
wolfspraul: (kicad) the one documented in gta02-core should be fine
<wolfspraul>
1) plugging multiple pieces together is not going to work for end users. too much will go wrong, dongle lost, connector breaks, etc. costs are also too high.
<wolfspraul>
2) going via USB means up- and down-regulating to 5V, so the power consumption of this solution will never be competitive with an integrated solution, issues such as sleep/suspend modes, incoming calls/connections etc. make it even worse
<wpwrak>
Necrosporus: 2b) there may be other concerns regarding openness, particularly if the technology is heavily proprietary or patent-encumbered in all other regards. in my opinion, that doesn't have to be a knock-out criterion, but it means that the effort put into this should be very limited.
<wolfspraul>
3) if we are licensing (buying) proprietary technology, such as the one neatly outsourced in your USB dongle, then I don't understand what the point of our project should be. Why not just use a really nice and integrated proprietary product right away?
<wpwrak>
Necrosporus: 3) must fit into the system. mechanically, electronically, in kind of use.
<wpwrak>
wolfspraul: for 1), a full-sizes A port would help at least with that tiny BT dongle. that's about the only sensible path to BT i can see at the moment.
<wpwrak>
on my fishing expeditions into RF-land, i also keep on looking for any BT or WLAN bycatch, but so far, i haven't found anything that's usable and that comes at least with a basic data sheet :-(
<wolfspraul>
yes but full-size A is big, we looked at it it wouldn't fit, plus I am just not motivated to spend my time integrating proprietary technology. if someone else wants to do that they are more than welcome :-)
<Necrosporus>
Internal micro PCI
<wpwrak>
wolfspraul: about 3), i think it's okay of we accept it as a temporary solution. we just should make sure a route to more openness exists.
<wpwrak>
wolfspraul: i'm still unconvinced big A wouldn't fit :)
<wolfspraul>
I knew you would say that :-)
<wolfspraul>
ok, for the server
<wolfspraul>
we use the 02-core kicad then, and --plot
<wolfspraul>
keep things simple
<Necrosporus>
It's possible to sell NanoNote without modules
<wolfspraul>
so the only thing is how we find the root .sch
<wolfspraul>
is there normally only 1 .pro file per project?
<wolfspraul>
some people don't seem to use .pro files at all
<wolfspraul>
why that?
<wpwrak>
wolfspraul: (wpan) the next steps: draw the remaining footprints (just a few qfn and solder pads missing), make the schematics of C8051F326 and AT86RF230 reference design (the latter with external balun, according to application note AVR2004), plus TI's antenna
<wpwrak>
(wpan) make the layout. build it. test voltage levels. test oscillator frequency and stability. move USB stack over from IDBG (this wants a clean solution, but not sure if i'll get to making it now or if i'll just branch and jump) add USB to SPI processing, along with a few "high-level" operations due to limited response times allowed by the chip
<wolfspraul>
he, sounds great
<wolfspraul>
what about the .pro ?
<wpwrak>
(wpan) read chip's ID registers from host. tune the crystal from host. crank up RF RX and scan the spectrum for signal strength (f/power plot, kinda spectrum analysis). compare with wifi activity and what the USRP sees.
<wpwrak>
(wpan) finally, crank up the RF PA and establish communication. watch the spectrum with the usrp. then make an f/tx-pwr/link quality plot. play with distance and location. decide if the thing is of any use :-)
<wpwrak>
(wpan) if yes, examine competitors to see any if their chips are better. e.g., ti have a similar one that's a bit smaller. almost all the ieee 802.15.4 chips have good documentation.
<wpwrak>
wolfspraul: (.pro) not sure if you can get away without even having a .pro file. you can of course choose not to share it
<wolfspraul>
wpwrak: well you tell me, you know kicad much better. we can just use the RootSch= from the .pro file, and work that way
<wpwrak>
wolfspraul: (.pro) generally only one .pro per project. e.g.,. gta02-core has multiple .pro files, but that's because it splits the kicad project into derived (by machine) projects.
<wpwrak>
wolfspraul: i think you're fine if you just go for the .pro, yes. any simple project will have just one. complex ones like gta02-core will have their own extra build rules anyway.
<wolfspraul>
yes but that's the point, make a system that is flexible enough for kicad in general
<wolfspraul>
I can see the value of scripting on kicad files for several purposes, not just to create png/pdf schematics
<wpwrak>
wolfspraul: run a project-provided shell script ;-)
<wolfspraul>
btw, the wpan-atrf.pro does not have the RootSch=
<alcy>
just order the nanonote from IDA :)
<wpwrak>
wolfspraul: scripting in kicad is still a long way. one obstacle that the guys in charge really like their gui. so "scripting" for them usually means "clipboard"
<wolfspraul>
painful
<wolfspraul>
hmm
<wolfspraul>
still not happy don't seem to find a nice and clean and quick solution for the schematics :-)
<wolfspraul>
either it's nothing (we have that already) or so general that it will take forever to implement (like running free scripts)
<wpwrak>
wolfspraul: oh, rootsch was supposed a tag in .pro ? didn't even know :) you can simply go for the same name
<wolfspraul>
alcy: great! thanks!!
<wolfspraul>
I hope you will not regret later, and this will lead you to an interesting development path. It should!
<wolfspraul>
and the device should also be functional day-to-day today!
<wolfspraul>
hopefully
<wpwrak>
i think you can just try the simple case and for n in *.pro; do sch="`echo "$n" | sed 's/\.pro$/.sch/`"; [ -r "$sch" ] && eeschema --plot "$sch"; done; even with gta02-core,
<alcy>
should arrive in about 3 days. thanks, wolfspraul. lol, I hope so too.
<wolfspraul>
ok
<wolfspraul>
so for every .pro. you take the corresponding .sch and plot that
<wolfspraul>
you think that will normally be the right thing in a kicad tree
<wolfspraul>
alright!
<wolfspraul>
:-)
<wpwrak>
wolfspraul: (wpan) if things work well on the hardware side, the software side needs a look. e.g., see how the zigbee and 6lowpan projects are doing.
<wolfspraul>
yes it all sounds SUPER INTERESTING
<bartbes>
somehow that sounds as if it is sarcastic to me
<wpwrak>
wolfspraul: (.pro) i think such a strategy would even work on gta02-core. one gotcha: if your schematics are in a different directory from the root, that would be harder to find. but then, i'm not sure such a setup would be very common. just have a means to turn it all off if it doesn't work.
<wpwrak>
bartbes:Â Â is the use of the word "super" a giveaway ? :)
<bartbes>
nice how the 2 of you have different conversations at the same time
<bartbes>
doesn't it make sense to.. you know, make sense?
<wpwrak>
bartbes: there is admittedly the small problem that ieee 802.15.4 is a bit of a niche at the moment. also, it's advertized as "personal area", with correspondingly small distances. but we'll see about the latter. some of the chips already seem to be some 20+ dB better than the standard requires.
<bartbes>
so ehm
<bartbes>
what do you connect the 802.15.4 to?
<wolfspraul>
bartbes: not sarcastic
<wpwrak>
bartbes: ZigBee, RF4CE (sp? now merged into ZigBee), 6loWPAN, anything new you come up with. it's mainly things like snesor networks at the moment, but i guess rf remotes and such will follow before too long.
<bartbes>
well, how do we get internet connectivity from it?
<bartbes>
I doubt it is somehow compatible with 802.11
<wpwrak>
bartbes: 6lowpan should take care of that
<bartbes>
but it needs a separate router, doesn't it?
<wpwrak>
bartbes: by the timne the ya comes out, the ipv4 address space should be exhausted already, so we'll be very trendy with an ipv6-centric solution ;-))
<bartbes>
heh
<wpwrak>
yes, need some kind of router
<wpwrak>
(ipv6) you could of course tunnel ipv4 over it or maybe even tweak the protocol to support ipv4 natively. so in case ipv4 refuses to die, we can handle it :)
<wpwrak>
rafa: he does sound as if he's saying that avoiding usbboot would also solve general usb problems.
<larsc>
wpwrak: the rom is only used to load another binary to do the real work.
<bartbes>
wpwrak: not to me, I agree with rafa, of course this is only interpretation though
<rafa>
we can ask him :D
<n0153>
Yes, I know about this page, but this page refers to max_posedon, and he adviced me to write to Yi@sharism.cc
<n0153>
I already do it, but there are no answer
<wpwrak>
larsc: ah, so it's rom+usbboot or rom+xbboot+kernel.
<larsc>
more or less
<larsc>
yes
<wpwrak>
larsc: okay, then it makes sense. thanks !
<larsc>
the problem is that the binary loaded by usbboot is not very fault tolerant. thats why we see so many problems with flashing
<wpwrak>
n0153: maybe ping her again ? or mail wolfgang
<n0153>
I ping her twice already
<n0153>
How to mail Wolfgang?
<wpwrak>
n0153: wolfgang@sharism.cc
<n0153>
Thank you! I will try
<wpwrak>
n0153: good luck ! importing things can be fun :)
<wpwrak>
larsc: so usbboot and xbboot have a different codebase ? (sorry for the noob questions. still only getting my bearings.)
<larsc>
wpwrak: yes
<larsc>
xbboot is quite minimal
<wpwrak>
i like the "let linux do it" approach better anyway. no use in replicating the same functionality over and over again.
<wpwrak>
wolfspraul: btw, did that story of the ben lost in buenos aires have a happy ending ?
<DocScrutinizer>
/kick HtheB
<wolfspraul>
wpwrak: hah, not really. turns out the guy outsmarted our payment system and used a stolen credit card :-)
<wolfspraul>
meanwhile the Ben is back in safety in Hong Kong, of course we lost some money for shipping and all the attempts at delivering the ben
<wolfspraul>
could thing we didn't succeed
<wolfspraul>
unfortunately it's too much trouble to bring forward the case legally, so he gets away with this nonsense
<wolfspraul>
sometimes that's real life :-)
<wolfspraul>
in about 900 Bens sold we had 2 such cases now
<wpwrak>
wolfspraul: wow. what a bastard :-(
<wolfspraul>
well yeah, real life I don't mind. we need to strengthen our systems and we do :-)
<Textmode>
only two in 900 doesn't sound so bad.
<wpwrak>
i should ask around if anyone knows some scary folks. send them over to show him how far the arm of the chinese mafia reaches ;-)
<Textmode>
not that it makes it any less annoying, but...
<wolfspraul>
Textmode: we could have caught both of them, our mistake...
<wpwrak>
Textmode: what's sad about it is that, because the shipment was stuck in customs and the guy couldn't be reached by phone or by mail, wolfgang traveled all the way from china to argentina to tell him personally. such a fine customer service wasted on a scoundrel.
<Textmode>
ouch...
<wolfspraul>
ha ha
<wolfspraul>
Werner exaggerates a little
<wolfspraul>
I traveled for the famous Casa Almesberger asado
<wolfspraul>
plus a nice soccer game
<wolfspraul>
but yeah, of course, I did follow up and try to give this guy good customer service
<wolfspraul>
as I would do with anyone who buys a Ben NanoNote :-)
<wpwrak>
n0153: so there is hope ;-)
<Textmode>
huggles wolfspraul
<rafa>
Textmode: Wolfgang exaggerates a little
<rafa>
it was not a nice soccer game
<Textmode>
heh
<bartbes>
is there a way to run gdb on the nanonote?
<bartbes>
(or a gdb server, and then gdb on my comp)
<bartbes>
because "Segmentation fault" by itself doesn't provide me with too much information
<rafa>
bartbes: could you install it? opkg install gdb?
<bartbes>
really?
<bartbes>
wow..
<rafa>
I think so.. at least I do
<bartbes>
so ehm, where do I get the package from
<bartbes>
I never knew how to get packages without compiling
<rafa>
do not you have opkg?.. I mean, I was thinking that openwrt uses opkg as well. opkg udpate; opkg install gdb
<bartbes>
really?
<bartbes>
sets up internet access on his nn
<rafa>
bartbes: I am not sure, I use jlime and I have gdb on repo to install with opkg. IIRC I read from other guys here using opkg in openwrt
<bartbes>
I first had to run update
<bartbes>
but then it worked
<bartbes>
hmm
<bartbes>
corrupted stacks
<bartbes>
nice
<bartbes>
ehm.. weird
<bartbes>
ah it doesn't support mips
<bartbes>
(remote debugging)
<bartbes>
builds the gdb in the toolchain
<bartbes>
come on toolchain.. build faster
<rafa>
bartbes: is not gdb on repo useful?
<bartbes>
rafa: I need the gdb from the toolchain (the one that runs on the dev computer)
<bartbes>
or at least, if I want to use gdbserver
<bartbes>
because the standard gdb on my comp doesn't support mips
<rafa>
bartbes: ??
<bartbes>
basically
<bartbes>
I'm running the application on my nanonote
<rafa>
bartbes: standard gdb?.. did you install gdb from opkg?
<bartbes>
but debugging on my computer
<rafa>
bartbes: ah.. okey. it sounded me that you wanted to debug on nn
<bartbes>
I semi-do
<bartbes>
but the gdb on the nn was easy to set uo
<bartbes>
*up
<bartbes>
as you said, opkg took care of that
<bartbes>
now I'm setting up my comp to do remote debugging
<bartbes>
ehm
<bartbes>
is the nanonote little endian or big endian?
<bartbes>
oh
<bartbes>
mipsel --> little endian
<wejp>
little endian
<bartbes>
I think that caused my segfault
<bartbes>
because a platform-dependant check (not directly related to the endianness btw)
<bartbes>
didn't include a header
<bartbes>
the endianness is simply because I'm now circumventing the check
<qi-bot>
[commit] Mirko Vogt: config: preselect qt4-packages which require nptl-support, since uClibc 0.9.32 with nptl support got merged over, preselect package <NanoMap> http://qi-hw.com/p/openwrt-xburst/73e7e7a
<qi-bot>
[commit] Mirko Vogt: check for <CONFIG_UCLIBC_VERSION="0.9.32"> instead of <CONFIG_UCLIBC_VERSION_0_9_32=y>, as CONFIG_UCLIBC_VERSION_*-booleans are only exposed, when <DEVEL=y> http://qi-hw.com/p/openwrt-xburst/d770d66
<kristianpaul>
wpwrak: hello:)
<kristianpaul>
wpwrak: i read you are planning do some stuff using SPI port
<kristianpaul>
are yuo using a native implementation of SPI in the ben, ifis truw, where is it?
<kristianpaul>
or planning bitbanging*
<kristianpaul>
(there is mail with this question too)
<viric>
Sorry; does anybody now what doom implementation people use in the nanonote?
<bartbes>
prboom?
<viric>
I don'tk now
<viric>
is it?
<bartbes>
the doom I have is prboom
<bartbes>
that I know
<bartbes>
and that is the one mentioned on the wiki
<bartbes>
by.. tuxbrain2 iirc
<mth>
bartbes: CongoZombie's doom port is also popular
<mth>
on Dingoo, at least
<viric>
ok
<viric>
thank you!
<viric>
SDL can be built so it runs on framebuffer, and it does not require X, right?
<mth>
yes, SDL can run on the frame buffer
<mth>
probably that driver is included in the default SDL for NN
<zear>
mth, i already ported czdoom to the nanonote
<zear>
and there's also yet another port of prboom to it
<zear>
i don't know if this is exactly the right place to ask this question, though it is about the qi mailinglist, so i guess it's related
<zear>
recently i receive more and more mails from the mailing list that are splitted from the main "thread", and go directly to my main mailbox rather than being sorted as mailinglist mails
<zear>
i never had to deal with any other mailinglist, so pardon this silly question
<zear>
this is quite annoying, because now my mailbox is spammed with mailinglist mail
<tuxbrain2>
zear, list disscusion y development has been merged, you have to reaply the rules to this new mailing list.
<zear>
ah
<zear>
that could explain it
<viric>
Does the SDL for the nanonote (let's say, in openwrt) use simd instructions?
<qwebirc54507>
hi
<mth>
viric: afaik the only SIMD instruction using program is Ingenic's MPlayer port
<viric>
I tried to build it...
<viric>
but it's quite coupled with *their* toolchain
<mth>
yes, it uses some awk script to preprocess the asm
<viric>
well, that would not be a big trick
<mth>
also, their MPlayer assumes it runs with their framebuffer drivers, which has some additional ioctls
<viric>
I'm afraid they have patched binutils
<viric>
Ahh really?
<viric>
ouch.
<mth>
they haven't patched binutils, but that is what they should have done
<viric>
We would have the source for that, right.
<mth>
mxu_as is the awk asm preprocessor
<mth>
it's in Ingenic's SDK
<viric>
wow.
<viric>
Did you see that mplayer running?
<viric>
does it perform much better?
<mth>
I haven't tested it myself, but I heard it did run better
<mth>
I don't know how much of that is due to the SIMD instructions and how much due to the IPU
<viric>
with the usual mplayer, does people manage to play anything?
<mth>
zear: did you compare MPlayer versions?
<viric>
I tried to put a mp4 video, and it went at 0.4 fps iirc
<viric>
(it was debian mplayer)
<viric>
I've been trying to build my own, but the build halts at a gcc file compilation *without any gcc message
<viric>
*
<mth>
if you transcode the video to 320x240, it should play smoothly even without hardware acceleration
<viric>
ah great
<viric>
mth: what codec you have in mind?
<mth>
years ago I used divx on a 200 MHz ARM
<viric>
the divx implementation in ffmpeg?
<mth>
MPEG2 is easier to decode, but results in larger files
<mth>
I transcoded using mencoder, I think it uses ffmpeg
<viric>
I meant the decoder
<mth>
I don't know, whatever mplayer selects by default
<viric>
it was mplayer on the arm? ok
<mth>
Sharp Zaurus
<viric>
ah I've heard about it
<viric>
So, I can't build mplayer! weird.
<viric>
It looks like a gcc bug
<mth>
I've never seen gcc fail silently though
<viric>
It's the first time I see it
<viric>
make says 'error code 2'
<mth>
I've seen intenral compiler errors or complaining about legal code
<viric>
always the same source file.
<viric>
In fact it looks like the whole 'make' dies at once.
<mth>
we even had miscompiled code once, but it did produce something then
<viric>
weird
<mth>
Make doesn't say which rule it was working on?
<viric>
I don't see even make.
<viric>
(I just noticed)
<bartbes>
can you.. paste the output?
<viric>
Mmm I think I tracked that to something I use in-between
<viric>
sorry for the noise
<viric>
I got the gcc failing message:
<viric>
codec-cfg.c: In function 'parse_codec_cfg':
<viric>
codec-cfg.c:546:24: error: 'builtin_video_codecs' undeclared (first use in this function)
<bartbes>
that seems like it's missing an include dir and/or header though
<viric>
I'll check
<zear>
mth, i haven't been really messing with mplayer for the nanonote
<zear>
well, tried to both run the dingux binary, and compile it myself, in both cases it wasn't working
<zear>
but that was on a very early version of firmware for the nanonote