Log from #neuros IRC channel

<crweb>

thats not something we have access to is it?

probably means NMS

<crweb>

ah

dunno, i just guessed that

or maybe may1937's work

<crweb>

i never grasped what may was doing

<crweb>

without the dsp compiler..

assembler

<crweb>

ah

<crweb>

either way, not much you can do is there?

[04:13]

well, having some way to load code into the DSP and have it execute that would be useful to some people

[04:13]

not me, granted

[04:13]

but still useful

<crweb>

but the code the dsp would execute would have to be assembled for the dsp

yeah, or by someone who has the compiler

<crweb>

which we don't..

<crweb>

i didn't think neuros did either

they do

have the compiler, i mean

<crweb>

oh

<crweb>

well then..

<crweb>

wonder what I have to do to get that..

they had it for the old N2, and the dsp is the same

crweb: you have to shell out some thousand dollars

<crweb>

may did?

he does assembler, that is supported by gnu tools

<crweb>

so there is a ts54x assembler

yes

<crweb>

well then

<crweb>

i didn't know that

<crweb>

I can do that

<crweb>

basically.

<crweb>

very basic

<crweb>

i could do hello world

<crweb>

heh..

<crweb>

square function

<crweb>

i^x

[04:18]

<crweb>

very interesting..

[04:18]

<crweb>

wonder how hard it is to port asm to asm

[04:18]

what do you mean ?

[04:19]

<crweb>

each assembler has its own format

[04:19]

<crweb>

wonder how hard it is to port assembly source to the 54x's assember's source

[04:20]

well, it probably has also lots of specific instructions you will want to use, otherwise there's not much gain in a dsp after all

<crweb>

dang it, its a C54x

yep

gdspsim.sourceforge.net

<crweb>

ts was the model we were using in class

[04:22]

<crweb>

ok, so do you know how we load the program?

[04:27]

no, there's what may1937 was working on, i think

[04:27]

<crweb>

ah

<crweb>

i've got to learn how to use nmsd

or you can use that simulator in the link i posted

<crweb>

i dont understand why it is a program linking.

<crweb>

yeah, have the simulator already

<crweb>

err, why its a program running

<crweb>

i know that its what I asked for, but i don't understand how they did it

[04:29]

<crweb>

so i make a socket to connect to it, and then tell it a file?

[04:29]

<crweb>

or do i pipe a file to it..

[04:30]

i have not looked into the new nms. i think you can still tell it to play a file and be done with it

[04:31]

<crweb>

do i specify a type or does it handle that to?

[04:31]

<crweb>

i honestly never did look at it at all

[04:31]

should handle it, the old one did, IIRC

[04:34]

<crweb>

hah, i left my 442 on recording osd output all night

[04:34]

<crweb>

filled the 40gig drive

[05:15]

<crweb>

nerochiaro: actually with mdev, I hope to not have to create device nodes ;)

[10:34]

crweb: you have been dead for 3 hours. go away

[10:35]

<crweb>

ping

[10:35]

<crweb>

interesting..

[10:35]

<crweb>

was crweb here?

[10:36]

<crweb>

some how my userlist didn't update

[12:22]

<crweb>

new AMD Live! all in one

[12:22]

<crweb>

neuros dev wins pricewise, geez

[15:20]

hi everybody..! has anyone tried to use eabi ?

[15:20]

<eigma>

shirour: hi :)

[15:24]

eigma: hi again..!

[15:30]

shirour: eabi ?

[15:30]

nerochiaro:yes! have you ?

[15:31]

shirour: i was asking what is eabi, actually

oh i see

shirour: no, never even looked at that, sorry

[15:35]

nerochiaro: ok, thanks..! it's built-in in 2.6.16+

[15:36]

shirour: are going to try that ?

[15:37]

maybe

[16:04]

oh wow, i'm trying .54 now, some pretty neat improvements i see. cool

[16:04]

it was some time i didn't try new versions

[16:14]

<shirour->

eigma - sorry, net problems.

[16:15]

<shirour->

eigma: got to sleep anyway... bye

[16:17]

<eigma>

shirour: bye

[16:17]

<eigma>

actually, you still there?

[16:18]

<shirour->

eigma:yes

[16:18]

<eigma>

firmware update succeeded.. building my special initrd right now, but I don't think I'll have time to do everything I need

[16:19]

<eigma>

oh well.. sometimes real life takes precedence over engineering :)

[16:20]

<shirour->

eigma: hehe..! ok! good luck!

[16:23]

<eigma>

thanks, good night

[16:23]

<shirour->

eigma: bye

[17:56]

<crweb>

nerochiaro: hey

[18:20]

crweb: hey

[18:21]

<crweb>

how goes it

[18:21]

actually i forgot what i msg'd to ask you ;)

[18:21]

<crweb>

sweet, well, i forog the answer

[18:22]

<crweb>

err.. forgot even

[18:22]

i just upped to .54 (without flashing). it was really a long time since i last updated, it looks better

[18:22]

lots of nice touches

[18:23]

<crweb>

yes, it is a lot better

[18:23]

<crweb>

now that I am not breaking it

[18:23]

heh, glad that's resolved in the end

[18:24]

<crweb>

yeah, and it appears we only lose about 1mb to imem

[18:25]

you tell that from the fact you tried 63M and worked, or from actual figures in the code ?

[18:25]

<crweb>

that it just works

<crweb>

I have filled the ram, with the codecs loaded

<crweb>

granted, i haven't played media

heh, that's actually the test that you need to run

<crweb>

yeah, soon as I figure out nmsd

[18:27]

want a quick piece of code to play a video ?

[18:27]

<crweb>

oh that would be lovely

[18:27]

1 sec

[18:29]

pastebin.ca i have not tested this past .35, might need some tweaking. i'd try that with .54 but my env is in flux right now

[18:30]

<crweb>

I had the hardest time getting the build to go today

[18:30]

which build ?

<crweb>

the whole thing

from trunk ?

<crweb>

come to find out it was dev node permissions

<crweb>

i dunno why they were wrong, everything root:root and -rw-r--r--

<crweb>

i had to set to -rw-rw-rw-

isnt' that how they're supposed to be (the second set you mentioned)

?

<crweb>

i believe so

<crweb>

some of them

<crweb>

all the tty's were crw-rw-rw- already

<crweb>

but null and console werent

btw, i mailed Gao, he said he will probably review the input stuff by end of week

<crweb>

cool.

hmm, some weirdness of the rootfs setup script. were you building from scratch, or from previous build ?

<crweb>

I wrote a input driver to read the neuros codes

<crweb>

from scratch

<crweb>

i never build from last build

wrote it as exercise ?

<crweb>

yeah, it was 18 hours of exercise too

<crweb>

I was actually learning how to write a QPlugin

[18:35]

<crweb>

the ir code took like 4 minutes to cp and paste yours from your input/event code

[18:35]

yeah, i imagine there's other machinery involved. the code to read from the irrtc device was pretty easy

[18:35]

<crweb>

i an now have keyboard input from any device you can imagine though ;)

what do you mean ?

<crweb>

I even just wrote a keyboard scripter

eh, all exercise

<crweb>

keyboard input from a file :)

<crweb>

I hope they get usb fixed soon though

to do usb hid ?

<crweb>

so we can use keyboards and mice

<crweb>

yes

what's broken in that ?

<crweb>

its going to save us a lot of trouble

<crweb>

that PIPE(0) thing..

<crweb>

I can't remember, someone was working with it though and hacked it to work

<crweb>

but not a good hack

but did you try the hack that Ycros did sometime ago ?

maybe it's what you are talking about

<crweb>

yeah, i think it was him

<crweb>

he implemented an empty PIPE_INTERRUPT

<crweb>

its actually suppose to do things

<crweb>

though I don't know what

and xiamen folks said they would fix it ?

<crweb>

no, but hopefully they will

<crweb>

it would open a very wide range of usb devices to us

if no one asks, i hardly imagine them getting the idea it's broken

[18:40]

<crweb>

last note about it was Ycros and I's thread

so they at least know the problem is there

anyone know if my OSD is dead?

my screen is simply the orange (eye and ear) screen

<crweb>

rastyk: well, let me look real quick

hehe.. i realize.. at least I put in the info real quick

and beat you to it :)

<crweb>

yeah yeah

<crweb>

i tried

<crweb>

no its not dead

<crweb>

what have you done?

<crweb>

nfs booting? just starting up with newest firmware?

well, loaded osd-3.27-0.49.upk, booted OK, played two .asf files that were on a usb-connected LaCie hard drive

<crweb>

so.. that doesn't sound dead

went to recording option, the cursor was in the TV (first line) and i hit the left cursor butt

on

screen just went blank. left for awhile, nothing. Tried the right button, back back, home.

<crweb>

nerochiaro: perfect example of why we need same version flash. If you can't read pal, and you oops make it pal, what do you do?

no response. POwered off and back on again, All i get is the orange "eye and ear" screen (the first one that comes up ) no movement

crweb: there should be some override for that. but i get your point

hope that information helps to explain what went wrong

<crweb>

hmmm.. if only someone in china was here for that.. where was that the other day when I needed it

<crweb>

rastyk: nope

<crweb>

:)

<crweb>

rastyk: do you have a serial cable?

that's the one that probalby came in the box, correct? let me check, brb

[18:46]

<crweb>

i dunno if it does or does not come in the box

[18:46]

it should

ok i checked and I have one, but there is a bit of a problem

<crweb>

no serial

it's a db9 male. My computers are all db9 male

<crweb>

oh, there should be a nullmodem connector

<crweb>

female to female

[18:48]

hmm I didn't see it

[18:48]

I only saw a male to a stereo-type of a cable at the other end (to plug into the other side)

[18:48]

I'll go check again though

[18:49]

<crweb>

it should be there

[18:49]

it's in the other room

[18:49]

<crweb>

its worthless without

[18:50]

<crweb>

nerochiaro: your code assumes that I have cooler and nanox

[18:50]

crweb: cooler, nanox shouldn't be needed

<crweb>

theres an aweful lot of widget talk in the code you pasted

crweb: eek, i probably left in some widget tests. sorry

<crweb>

ok

from IsProbeModulesDone onwards is media

and also before AppCreate

[18:52]

the stuff in between is widgetry

[18:52]

<crweb>

leave appcreate?

[18:52]

i think you can remove them. also the two calls at bottom

<crweb>

fb.h application.h

away

both

*shrug* I have an av/cable, a ir blaster cable, a db9 cable (that was mentioned)

also linux/fb.h

and also rtc.h isn't needed

i left lot of cruft in there, sorry

<crweb>

thas ok

<crweb>

i'm going to needs nms though

it's accessed through cooler

if you talk about the headers, yes

in other news, i fucking hate the script that copies and installs stuff to the rootfs even more

<crweb>

i aswell

<crweb>

you know how long i've been trying to figure out how to get /etc/networking into the build

you never know if everything you added is reinstalled on full rebuild

[18:56]

i can imagine

[18:56]

<crweb>

I've spent more time with that then writing the damn thing

[18:56]

<crweb>

i probably don't need signal handler

[18:58]

don't recall why i put that stuff there too

[18:58]

it's a bit my testing garbage dump

[18:58]

crweb: say, if you change a file in /scripts/targetetc and want it to be updated in the acutal fs, what the hell do you call ?

<crweb>

./setup-rootfs

ah, right

<crweb>

nms needs cooler

<crweb>

so much for it being a stand along player

<crweb>

err server

i keep thinking i need to make in bsp/rootfs

crweb: cooler is the lib used to access the server

<crweb>

i'm trying to build the NMS server

<crweb>

it needs cooler

<crweb>

thats backwards..

crweb: a bit. it probably also needs it for other stuff. cooler is a bit of the "all code that it's not osdmain goes here" kind of support library

<crweb>

well, shit

<crweb>

cause to get cooler, i need nanox

<crweb>

and to need nanox, means that I lose a ton of space for something I'm not going to use

<crweb>

whats the point of making it a server if it needs everything else

i can see what you mean. i'm making a nanox-less version of cooler, but it's not anywhere near ready right now

crweb: anyway, you can do this

crweb: go into cooler and strip off the widgetry stuff

crweb: it's not actually all that hard

<crweb>

nms is opensource right..

<crweb>

and it is used to access the codecs/kernel modules

yep

linux-r3-main-app/nms

<crweb>

so really.. It shouldn't be to hard to rip out something that just plays files

<crweb>

I have my own libs for reading id3 tags and such

yep, i don't think it will be that hard

by the way, setuproofs didn't do it

<crweb>

yeah, was afraid of that

<crweb>

this is why i rm -rf

<crweb>

cause make gets stuck thinking it already installed

<crweb>

you have to rm -rf rootfs/* EXCEPT for the make file

oh bollocks. thanks

<crweb>

yeah, you HAVE to generate

<crweb>

it pisses me off

actually you need only to rm -r rootfs/default

but yes, it's damn annoying

<crweb>

same thing

mostly

<crweb>

i remove the tmp too

<crweb>

cause I like stuff clean

<crweb>

but yeah