Log from #neuros IRC channel

daurn: you made that script some time ago to pick up the flv url from the video page in youtube, right ?

<daurn>

ages ago :P

let's see. if the OSD requests guide, and tells what timezone OSD is in, then the time can be fed back as a part of guide info. I suppose that's ntp alright.

yeah, but in fact isn't it just enough to find the "<object" and then pick up the source url from that '

?

[08:37]

<daurn>

all you have to find is the bit after &t= in the source

[08:37]

vmarks: too bad guide info gets in weekly or at most daily, ntp can update more frequently than that

[08:38]

daurn: ok

[08:38]

<daurn>

nerochiaro: i made mine into a library of sorts (never finished it)

[08:38]

<daurn>

you can ONLY grab that part if you want

[08:39]

we'll probably pick the idea and distill it into the youtube app in C

[08:41]

nerochiaro: why would you need to update more frequently than that?

vmarks: if it loses one minute a day...

is time drift on OSD that big a problem?

and if so, let it get guide nightly?

it's not that bad now

I mean, there's no reason to hammer an ntp server, is there?

vmarks: probably better to find the cause of drift and fix it

yeah, that's probably sensible.

vmarks: every few hours is hardly "hammer", i'm more concerned by adding the ntp stuff to osd will use even more space

so use the guide as the means of setting clock rather than NTP.

and do guide nightly, or weekly-and-fix-time-drift, right?

[08:45]

you'd want to get the guide daily for last minute tv schedule changes and that would most likely be sufficient to sync time

[08:45]

also i'm not exactly sure how ntp works, but it does something to compensate for the time it takes to reach the ntp server and back, no ?

[08:46]

otherwise time would drift a bit according to network lag

[08:46]

(i might be mistaken here)

it is so

nerochiaro: have you ever experimented with the arguments to the itfb mobule?

[09:01]

drawnout: one min. phone

[09:01]

nerochiraro: *nod*

[09:18]

drawnout: so, yeah, i messed with them

[09:19]

drawnout: turns out itfb sucks a lot

[09:20]

drawnout: basically the only way to change resolution is to unload the module and reload it passing a different resolution argument

[09:20]

among the few supported

[09:21]

nerochairo: *nod* I was just looking at that and trying to adjust the g_vmode.

drawnout: there's an header for itfb where it lists all the supported modes

nerochiaro: Is itfb going to be rewritten as a result of the SoC projects?

no

at least not that i know of

nerochiaro: where can I find the header? I'm currently using strings to find acceptable parameters

[09:23]

itfb.h i think it was added to the linux headers

[09:25]

there's also an fb driver from cowon that we can use as a base, it kinda worked when i last checked but crashed in some cases and i never had time to look at why. also didn't support some ioctls for transparenct and such, but these are trivial to add once we're sure it works for basic fb stuff

[09:26]

nerochiaro: *nod* I'm looking at this b/c I just got mine, and it's kind of scary looking on my tv, figured I'd try switching the module to ntsc-d1 output and hope for the widescreen res to work

[09:29]

if you could figure out the cowon driver, that would rock

[09:29]

so we can change res at will

[09:29]

*nod* I'll take a look at it after tomorrow. Research paper due tomorrow... crap gotta run

[09:30]

thanks for the help nerochiaro

[09:30]

k

[12:04]

cr-temp: crweb ?

[12:04]

<crweb>

:)(

[12:04]

<crweb>

hmmm.. :)

[12:05]

crweb: did you receive the soc meeting emails ?

[12:05]

<crweb>

yeah

[12:05]

<crweb>

i replied saying i needed it moved forward or back a hour

<crweb>

i have class at 11am cst

check mail again, we moved it. say if ok

and actually i never got your email where you said that, i just heard you mention it here the other day

<crweb>

perfect

please reply saying so then, so we're all set :)

<crweb>

18:00 is great for me

<crweb>

ok

nerochiaro: do you remember what things didn't work with the cowon fb driver?

drawnout: i only know it crashes on changes of resolution, but i never figured out why

<crweb>

still on the resolution thing aye?

<crweb>

;)

nerochiaro: 'k, that's a good enough starting place

crweb: no, drawnout is there

drawnout: maybe i have some source you might want to look at

crweb: got itfb to load hdtv-720p today

<crweb>

video out resolution isn't itfb resolution

<crweb>

just keep that in mind

crweb: that's what I'm still confused on, does something downgrade the signal to vanilla ntsc/pal?

<crweb>

drawnout: think of it as 2 video chips

drawnout: it's 2 layers. video and fb on top of it. composed togheer

<crweb>

drawnout: the video out from the dsp, and the framebuffer

<crweb>

drawnout: they both go into a single NTSC/Pal chip that combines the signal and outputs it to tv

crweb: so there's no hope of ntsc-d1?

<crweb>

unless you can hack the ntsc chip, i don't think so.

<crweb>

it is analog

<crweb>

we can't suddenly change it to digital

[12:11]

<crweb>

its a hardware thing

[12:13]

<crweb>

nerochiaro: do you pronouce your name like You Go?

[12:13]

*nod* my osd is always gonna look ugly on screen :(

[12:14]

crweb: nope, the "U" sound doesn't have the "y" part, i'm trying to think a word in english that has that sound

[12:15]

crweb: can't think of any though. but i don't get offended if someone pronounces it that way :)

[12:15]

<crweb>

hmm.. I'll ask dad when he comes over tonight

[12:16]

www.sonivius.com

[12:16]

crweb: yeah, if he's still good on italian pronounciation. years of speaking english do ugly things to pronounciation

[12:17]

Emmett: what the heck is that ?

[12:17]

Emmett: weirdness indeed

[12:17]

Me playing the accordion.

[12:18]

the base worried me. accordion is nice

[12:20]

<crweb>

i just got a new 64bit machine

<crweb>

well, not to new, athlon 64, 2.0ghz

<crweb>

but it does xvid encoding at about 80 fps

<crweb>

where my Pentium M 1.73 does 10fps

<crweb>

and my pentium 4 3.4 does 32fps

sweet

<crweb>

its really a slower cpu though..

drawnout: still there ?

[12:22]

<crweb>

this is all very strange

[12:22]

<crweb>

64bit can't make that much difference

[12:22]

crweb: just FYI, EPG stuff was shifted down in priority in schedule, but if you want to see the bit of db design i did, i can send you

[12:23]

<crweb>

yes, please do

[12:23]

<crweb>

I can start hacking up some of the db code atleast

[12:23]

<crweb>

i'm pretty good at that

[12:26]

sending out in a second

[12:29]

crweb: sent to your inbox. there's some crap in that, the doc is generated with a tool but it's outputing also a lot of useless stuff

[12:29]

you should get an idea anyway

[12:30]

<crweb>

alright

[12:30]

<crweb>

i got class then I'll take a close look

[12:31]

<crweb>

due to the design of sql3

[12:31]

<crweb>

i think we may need a EPG server

[12:31]

<crweb>

to handle multiple connections

[12:32]

<crweb>

i don't think you should have multiple programs connected to the same lite db

crweb: you can share connections, i think there's a locking system

in the library. i read that in the docs, iirc

hi, I have kind of a basic question. If you're recording, and the power is cut to the device what happens?

is the video lost?

[12:35]

GoatBoy: i think it's written up to a certain point

[12:36]

is the container format such that the video is recoverable even if it's truncated?

[12:37]

i'm not sure, but is that something that happens frequently for you

[12:37]

?

[12:37]

if I wanted a full-time in-car recording system, yes...

[12:38]

it's easier if the thing is just resilient to powerdowns as opposed to constructing some rube-goldberg shutdown mechanism

[12:39]

right, however i'm pretty sure that even with missing header info at worse case, the video data is there so you can recover it with some software

[12:39]

on a PC

[12:41]

k sounds reasonable

[12:45]

<pb101>

nerochiaro, are you familiar with NMS?

[12:45]

pb101: try asking :)

<pb101>

nerochiaro, OK, I can't get the server to play anything

<crweb>

nerochiaro: that would be fantastic then

crweb: what ?

<crweb>

nerochiaro: i've never looked into the locking. I'vei always had sqlite statically linked in my app so i didn't have to worry about it

[12:47]

<pb101>

NmsConnect() seems to work, but NmsPlay() just fails because of errors when sending the command.

[12:47]

crweb: i need to check, but i think i read it

[12:47]

pb101: which version are you on ?

[12:48]

<pb101>

nerochiaro, Nms is returning 1.0.2

[12:48]

no, of the source

[12:48]

trunk ? a tag ?

[12:49]

crweb: www.sqlite.org

[12:51]

<pb101>

nerochiaro, I got downloaded the latest version a couple of days ago, but I am linking againt the lib that came with the latest firmware.

[12:51]

pb101: are you using the VM then ?

[12:51]

<pb101>

nerochiaro, yes.

[12:53]

so wait, you are mixing libraries from different versions ?

[12:54]

<pb101>

nerochiaro, I am compiling on the VM, but the final code runs on the OSD.

[12:56]

<pb101>

my new compiled code talks to the nms server on the OSD, which I didn't compile.

[12:57]

pb101: well, let's say picking that library from the OSD and putting it in the VM is ok for now. what errors give you nmsd when you try to send the play command ?

[12:58]

nerochiaro: back now, had to grab lunch

[13:00]

drawnout: check out svn.neurosaudio.com

[13:00]

drawnout: maybe something in there can help you

[13:01]

nerochiaro: ok, thanks. I'll take a look

[13:01]

<pb101>

nerochiaro, ./../src/cmd-nms.c--GetAll: 94 retry timeout! followed with: CMD rcv error, Unable to send command, etc.

[13:02]

pb101: i'm thinking that maybe you should also pick up the updated header for the library you picked up from the OSD

[13:02]

<pb101>

nerochiaro, I do see nmsd running as a process.

[13:03]

<pb101>

nerochiaro, would the headers be stored on the OSD?

[13:03]

no, you need to pick them up from svn, from the same tag that your library was built from

[13:04]

<pb101>

nerochiaro, I will give that shot...thanks!

[13:05]

pb101: i'm saying that because with mismatched headers you might be using some wrong constant or something else that funks up the way you send down commands with the server

[13:06]

<pb101>

nerochiaro, Makes sense. By the way, whay you think of using server-side calls to play media falls and bypassing nms?

[13:07]

pb101: bypassing what ? the library ?

[13:07]

<pb101>

nerochiaro, bypassing nms.

[13:07]

i don't understand what you want to do, sorry. can you explain a bit better ?

[13:08]

<pb101>

nerochiaro, basically, instead of using the nms client library to communicate with the server, I would directly call the functions used by the server (SrvPlayFile(), etc).

[13:09]

pb101: what advantage would you have doing that ?

[13:10]

<pb101>

nerochiaro, I already have an RPC server (TCP/IP), which already handle multiple clients, locking, etc, so it seems redundant to interact with the nms server.

[13:10]

you have that on the OSD ?

[13:10]

<pb101>

nerochiaro, yes.

[13:11]

<pb101>

nerochiaro, it handle file upload, download, image displaying...I am just trying to add the ability to play media files (MP3, MPEG, etc).

[13:12]

well, ok, but if you're the only one using the server

[13:14]

<pb101>

are you referring to osdmain?

[13:14]

pb101: or anything else we might have in the future that needs to use it. actually i'm not even sure why that interface is exposed from the server.

[13:16]

<pb101>

nerochiaro, well, it's in the source code for the server...I don't know what you mean by exposed.

[13:19]

pb101: how would you call these functions ?

[13:20]

<pb101>

nerochairo, the client would send a request to the RPC server, and the server would call the appropriate functin to play the MP3, for instance.

[13:20]

pb101: call it how ?

[13:20]

pb101: the rpc server is an app. nmsd is another app

[13:21]

<pb101>

nerochairo, the rpc would bypass nmsd, and directy the function that nmsd is using right now.

[13:21]

<pb101>

directly call.

[13:22]

pb101: you mean that you compile part of nmsd into your rpc server application ?

[13:22]

<pb101>

nerochairo, yes, is that not OK?

[13:23]

no, ok, but then everytime something changes in neuros' "official" nmsd server, you need to take these changes and re-apply them to your application

[13:24]

<pb101>

nerochiaro, yes, I know...that's why I am asking about the pros/cons of using the NMS interface!

[13:26]

pb101: i don't think you need to worry too much about overhead of using the client library, and saves you lots of trouble. if later you see performance problems, you can think of optimizations.

[13:28]

<pb101>

nerochiaro, at the moment, I only want the nms client to work :) I could be using incompatible headers/libs, so I will reload what I have and try NMS again.

[13:28]

ok. try that and let's see what happens

[13:28]

<pb101>

nerochiaro, cool, thanks!

[13:58]

<crweb>

if you can think of optimizations, you can take them and put them right back in nms

[14:03]

<crweb>

you know honestly.. i think i'm just going to change all my clocks to UTC

[14:04]

<crweb>

the only thing i use CST for is class times( only have to convert once), and TV.. which i don't have

[14:06]

heh, interesting idea

[14:08]

<crweb>

i mean seriously, all you really use your local timezone for is day/night

[14:08]

<crweb>

but I don't use day/night anyway

[14:12]

haha

[18:39]

<pb101>

nerochiaro, u there?

[20:57]

zzz

[20:57]

quiet in here

[22:26]

* Johndesigner watches tumble weed roll by

[22:27]

* vmarks abuses JoeBorn's handheld.

[22:28]

bored?

[22:29]

sshfs#case@slug.local:/storage 7.5T 0 7.5T 0% /home/case/Music/

[22:29]

:>

[22:29]

how much music do you plan on having?

all of it it looks liek

before i divorced i was @500GB

you win

so heh, since it's one big interest of mine, it'll grow.

well, it's good to dream

i'm wetting my pants over a oqo too

but dunno when they'll hit the market.

vmarks, got any inside info?

on oqo?

yeah

no ideas.

the sony umpc looked more interesting to me, just because it's a core solo instead of whatever very old lower power intel is in oqo

[22:49]

the oqo is a via.

[22:49]

is the sony umpc on the market?

a via?

fuck that.

and while I suppose that's not bad, it's not for me.

the sony one has been available for a little wile.

while.

intel core solo 1.3 ghz. 1gb ram, 32gb flash or 40gb hdd.

[22:54]

interesting

it doesn't however inspire me.

silly I know :)

well, if it cost about 900 less, I would be more interested.

but oh well.

[22:56]

I like my laptop, but I was tired of dragging it all over the business trips.

[22:56]

just wait for the olpc

you're not serious.

...

why not?

the olpc is a novelty item at best.

[22:57]

oh yeah i can't wait to get one, ruggedized, small, personal powered

[22:57]

me too

hackable

try using inkscape/xara/illustrator on it

that would be troo

true

that would suck, but that's not what I would use it for..

is there any up-to-date guide than the derobert's guide for checking everything out of svn to set up nfs build environment?

it's targeted at developing nations. cool. however, something more up to modern chipset like the sony umpc or even that oqo, are going to mean that I can travel and do the more modern tasks I find that I have to do.

olpc is a fine idea - it's just not for me, for most of the stuff that I do.

good thing i don't do graphics

vmarks, i like the form factor of the htc phones

htc is a leader in that.

although I haven't seen the new one you're referring to

definetly getting one of their wm6 phones later.

you still have to get me the meizu m8

I guess I do owe you that.

:)

i know it'll run emacs and gcc that's about all i need

at this point, I'm either checking email or doing computation intensive stuff that wouldn't happen on a laptop anyway, so the olpc sounds pretty good to me

perhaps mplayer

gl2tosl2: sure, for your needs it works.

xcasex: you mean the htc shift device, right?

vmarks, any way. we're trying to present the viewpoints of graphics artist & bussinessman vs. coders/scientist

vmnyeah

vmarks, yeah

certainly there are market opportunities for both

different needs require different computers and even form factors.

agree.

fair enough

I think I might like a tablet pc though

i'll probably keep the tabletpc, but I need something "lighter" for other more mundane tasks. and an ultra portable fits the foot for me.

can't tell unless I get to play with one for a month though

and if all I had to do were email and could move everything else off to a server and attach to it, I could do an olpc. but I find when I travel, if I don't have everything with me and not dependent on network, I can be caught out when I need to show something or access a file.

I am not proficient enough with a tablet pen to make it work for me.

oh hmmm, i bet you can't even give a presentation off it