Log from #neuros IRC channel

<crweb>

i mean, that would work yes

ah, didn't think of that since i never had a problem with it

<crweb>

but what i want is a deb like scratchbox-libs

<crweb>

there are no docs on how to add libs

<crweb>

only how to build devkits

<crweb>

which isn't what we need

because it probably relates to how to build toolchains, i think

<crweb>

looked in there too

<crweb>

i cant find how it even decides what to copy from /scratchbox/host_shared

i would just ask in the #scratchbox channel. there are folks in there who helped me in the past.

(when they're on)

<crweb>

k

<crweb>

I'm done for tonight

best time seems to be from 21 in my time zone (GMT+1)

<crweb>

i hate html

<crweb>

it makes my hands hurt..

[02:29]

and yet you have "web" in your nick ;)

[02:30]

<crweb>

i used to be the admin for christianrock.com, christiansrock.com etc

[02:30]

<crweb>

admin/designer

[02:30]

<crweb>

that was way back in the day... .com boom, 1998-2000

ah, i was wondering what the cr was for

<crweb>

they are domain parked now

<crweb>

yeah, i'm stuck with the name

<crweb>

was a gentoo developer for a long time, and am crweb on all the forums, irc, etc

hehe, well, you can always say cr stands for something else ;)

<crweb>

it does

<crweb>

Christopher Richoven

who is that ?

<crweb>

or Columbus Richoven

<crweb>

my AD&D characters

<crweb>

heh...

<crweb>

ones a paladin, the other a cleric

* crweb was a dork

AD&D is sooo 1980, dude. you should move thm into WoW, these days ;)

<crweb>

hah

<crweb>

i can't play mmorpg's

<crweb>

addictive tendancies.

yeah, i know. i was joking. actually, i used to play lots of tabletop rolegames too

waaay too much of that, actually

[02:35]

<crweb>

yeah. me too

[02:35]

<crweb>

heh

[02:35]

most of the time i was GM (or the equivalent) though

[02:36]

<crweb>

high level paladins was my thing

[02:36]

<crweb>

and, i had a gm once let me be a dragon

ah, was forgetting. for html sores, just get one of these tiny converter scripts from some wiki-like syntax. for simple pages works the best for me

<crweb>

i may

lol, every GM who does that is a sign is getting overboard. that, and wishes

<crweb>

haha, well, i wasn't allowed to tell anyone

oh, and +something, attribute something swords

<crweb>

can't remember which dragon, but the one that can human morph

your private dragon fantasy ? ;)

<crweb>

it fit into the story

i imagine. i was just having fun at D&D stereotypes

<crweb>

yeah, we were a backwater group

[02:39]

<crweb>

i've never played with any other group, and they have never played with anyone else

[02:39]

<crweb>

so, there wasn't much gm showing off and that kind of stuff

<crweb>

i can't stand to play with the groups here at the university

<crweb>

they are crazy

kind of the same here, we were basically the geek kids in my town, and the ones where i routinely went for holydays in summer

strict roleplay code, nuts adherence to the rules kind of players ?

your uni players i mean

[02:41]

<crweb>

no, they are more like "look at my character hes so amazing" GM's compete to see how outragous they can be

[02:42]

oh, terrible. kills all the fun, even worse than rule-nazi

[02:42]

<crweb>

its a lot of showing off a piece of paper... that is... numbers written on it like they mean something

[02:42]

heh, "munchkins" they are called no ?

[02:43]

<crweb>

and of course, for all i know, they just wrote numbers in the boxes..

[02:43]

<crweb>

yeah

[02:43]

there's also a card game with that name, it pokes fun at D&D cliches and it's damn fun

[02:44]

to play i mean

[02:45]

<crweb>

ok, bed time

[02:45]

<crweb>

see you in +6 or so

[02:45]

good night. later

[02:46]

<crweb>

btw, i really like neuros-osd-toolchain

[02:48]

crweb: i think it's a good idea

[02:48]

crweb: oh, and also nano-x headers and libs please

[02:48]

<crweb>

not sure where to put them

<crweb>

due to linking problems

can you explain ?

<crweb>

would /opt/neuros/nms/lib & include work?

<crweb>

would /opt/neuros/nanox/lib & include

<crweb>

both those

<crweb>

then you can gcc -L/opt/neuros/nms/lib -I /opt/neuros/nms/include

it's too bad the source isn't arranged differently so all you need is a single include

-I/opt/neuros/include

<crweb>

sourcerror: thats not a good idea

not true

<crweb>

the seperation is good

you put them in sub directories.

and include in your code .... #include "nms/this.h"

<crweb>

exactly

<crweb>

thats not portable

<anders_>

Could we have neuros-env set $NEUROS_NMS_INCLUDE_PATH?

We do it for modules where I work it works fine.

<crweb>

sourcerror: you'd have to use "my" system

<crweb>

anders_: this is unrealated

anders_: it's $PRJROOT/../Neuros-Cooler/include

<crweb>

anders_: I'm removing the compile tools out of neuros-bsp

<crweb>

for a more native feel

if you arrange your source $MODULE/{public headers} and $MODULE/src/{source files} then everything works out without moving things around.

crweb: what you said about /opt sounds good to me. where's the problem ?

<crweb>

nerochiaro: i was just replying back to sourcerror

crweb: add also a pkg-config style script and voila'

see above

[02:54]

<crweb>

sourcerror: not talking about building neuros things

[02:54]

if all neuros stuff was arranged that way it would be great.

[02:54]

<crweb>

sourcerror: this is setting up cross compiling system wide, not using $PRJROOT, or neuros-env

<crweb>

sourcerror: again, its not for development of kernel stuff, or bsp stuff

<crweb>

this is for non-scratchbox general development

<crweb>

for now

shouldn't matter

[02:56]

<crweb>

you're worried about building modules

[02:56]

<crweb>

and building sources.

[02:56]

<crweb>

I'm providing prebuilt libs for linking

with their headers though.

<crweb>

developers dont' need to compile nms, or the kernel, to use nms or the kernel

<crweb>

i think neuros's dev setup does great

<crweb>

with the bsp and stuff

<crweb>

2 seperate systems

why do you keep saying nms, when you really should mean cooler ?

<crweb>

cause I don't really carea bout cooler

<crweb>

heh

but it's pointless. mns headers are just part of cooler

<crweb>

no they're not

nms-client.h ?

<crweb>

linux-r3-main-app/nms/include

Neuros-Cooler/include/client-nms.h

<crweb>

nms and cooler are not connected as much as we thought

way to go for duplication

$BASE/neuros/client-nms.h

<crweb>

nerochiaro: cooler/client-nms.h is for coolers connections to nms

<crweb>

nerochiaro: nms/include is for nms

crweb: it's for connections to nms

<crweb>

i was told they will be seperate

crweb: what else you do with an nms server other than connecting to it ?

<crweb>

thats what i mean

<crweb>

any app can connect using nms/include

<crweb>

for a cooler app to connect, you use cooler's client-nms

[03:02]

hmmm, you are right. but then it's stupid. doesn't make any sense to me, it's pointless redundancy

<crweb>

nerochiaro: cooler uses the server to connect to

<crweb>

if you use the server.h, you're using the libs, maybe?

and your app links in libnmc.so

...if it existed separately (doesn't yet right?)

crweb: so the server is built into your app

<crweb>

i dunno, i think we are all confused

crweb: ?

<crweb>

nerochiaro: well, the server is a .so, that you link to

<crweb>

i would think

the server should be actually an exectutable. a daemon

<crweb>

I still don't know why nms is a running daemon

the server is a standalone app 'nmsd',

what you mean you don't know ? it's a _server_

<crweb>

yeah, but why

when you want to talk to it you use the client which was part of libncooler.so

crweb: because multiple apps can connect to it, in theory

but should be pulled out into something linke libnmc.so

<crweb>

i've been using the lib directly

<crweb>

ah, right..

crweb: so you're not probably running it as a server. you're just building-in the server into your app

<crweb>

well, i keep saying nms because I plan on providing both the nms.so, and the client

sourcerror: you're right, i kind of agree

<crweb>

they are suppose to be seperate

<crweb>

from cooler

which only adds confusion, imho

<crweb>

and mgao said it would be very easy to take out of cooler

no, i agree on separation of cooler and a client library for nmsd, as sourcerror said

yea, look at the code it isn't much for the nmc (client)

but not in building-in the server into apps

<crweb>

nerochiaro: confusion? if you want to use the daemon, you link to the client, if you want to use the api directly you use the lib?

<crweb>

its just a matter of linking

i know, i just don't see the point of linking in into your app

i mean, let's move off that monolithic approach

<crweb>

maybe someone is building their own server

otherwise we could've left nms inside osdmain

we need lots of separate services to do more cool things in the future anyway.

<crweb>

i'm not being monolithic

<crweb>

i'm providing all the options

fair enough. all of them though, don't leave off cooler :P

<crweb>

no no, i'm not leaving anything out

:D ok

<crweb>

its just quicker to say neuros media system

<crweb>

than, cooler, nanox, osdmain, neuros media server, etc

ah... I always read nms as neuros media server and nmc as neuros media client

er, actually when you say nms one things s = server ;)

er, thinks

<crweb>

i treat it as a system

<crweb>

nmsd is the server

neuros media server daemon ;p

well, just to let you know you confuse people when you use it that way. i suggest you don't

but then, it's a suggestion :)

[03:11]

<crweb>

eh, i'm including all neuros libs and headers

[03:11]

<crweb>

nms (server) specifically first

[03:12]

<crweb>

then nano-x and cooler

[03:12]

<crweb>

i mean, sure i'm doing this for everybody. but i need nms/nmsd for my stuff, so its going to be done first.

hi

crweb, nerochiaro: did it work?

<crweb>

no, i think it is missing the pal/ntsc controls

<crweb>

which is why it didn't output

[03:14]

<crweb>

nerochiaro: it was probably outputing to the fb, just not the fb wired to the ntsc/pal

[03:14]

<crweb>

it was probably going to the lines that are split and not connected to anything

[03:14]

i agree

[03:15]

<crweb>

i think if we activate the ntsc/pal, it might just go ahead and show up

[03:15]

<crweb>

unless that chip specifically needs a specific resolution

[03:15]

to activate ntsc/pal we need to know what to tell to the hardware to do so, which is in the itfb source

which is closed

<crweb>

its also in the doc

which is closed

<crweb>

you dont' have it?

<crweb>

hmm..

should everyone have it ? is it available for free to the public ?

<crweb>

no

eh

<crweb>

you can dev with it, just can't release it

[03:18]

<crweb>

nerochiaro: nda isn't for source, it is only for the actual document

[03:19]

source which i don't have, too

[03:19]

<crweb>

you don't need the source

[03:20]

<crweb>

the doc tells you how to use the chip, you'd just modify the cowon driver

[03:20]

right, but where anyone gets this doc ?

[03:21]

<crweb>

sign with neuros

nerochiaro: are you working on the scheduled recording + guide

sourcerror: on the guide part mostly

just didn't have much time these last days

and also was distracted by this framebuffer thing

[03:27]

I tried to follow way back but wasn't sure where it headed

[03:28]

I guess there is the GUI part of the guide and the guide data itself

[03:29]

yes. the data you can gather from many sources. then you bring it on the osd and display it, allow people to choose the programmes they want to record and schedule them

[03:31]

I think the actual recording schedule needs to be its own service like nms but it actually contacts nmsd

[03:32]

this allows your guide to work alongside other schedulers like a remote scheduler or others

[03:33]

not a bad idea

<crweb>

it should be put into a db

<crweb>

and that is it

<crweb>

only touched when updated or read from

this frees it from the GUI

crweb: yeah, but right now you can record only when nmsd is running

crweb: with a scheduled recording service you need not

<crweb>

nmsd is always running

but it doesn't manage scheduled recordings, i think

<crweb>

it should read a db

<crweb>

right, i think it should

yes. i'm thinking of using sqlite now

<crweb>

it is the media server

the schedule service owns it and provides the abstraction

crweb: did it work well for you

?

<crweb>

nerochiaro: yes, it does

sourcerror: yep, but a library would be enough

sourcerror: no need to add another daemon

if you need only the abstraction

crweb: cool, i'll keep on that road then

[03:39]

either you give it life as a daemon or you require its responsibilities to be invoked by many other places. seems clean as a daemon

<crweb>

1 place

<crweb>

nmsd

<crweb>

which all apps have to use to access media functions anyway

not nmsd... too much crud in nmsd

<crweb>

crud?

its like small unix tools

we have cp and mv not cpmv

<crweb>

this isn't a tool

<crweb>

this is a system

sourcerror: what is wrong with a library that abstracts read/write to the scheduled recordings db ? then you schedule them with that lib from wherever you want. nmsd reads the db with that lib too, and does the actual recoridng

in my picture it is a service that multiple UI's can ask it to schedule recordings.

<crweb>

sourcerror: you don't need all that

it's not useful to have it as a service

<crweb>

you have the service, its nmsd

<crweb>

all UI's can schedule, by writing to the database

[03:42]

sourcerror: the act of "scheduling" rercordings, is just writing stuff to a file. you don't need a running program to do that

[03:42]

i mean, you don't need a running server

[03:43]

it may as well since an agent has to look at the DB periodically and tell IR blaster to change channels then tell nms to record

[03:43]

you may as well have this be a dedicated service

otherwise you such channel changing into what... nmsd?

s/such/suck/

<crweb>

why not?

<crweb>

you have to connect to nmsd anyway

<crweb>

it is a media device

<crweb>

the media server should handle media related functions

to change channels you don't need to connect to nmsd. you just need cooler. it has ir blaster functions in it

using a service (nms) doesn't mean you have to make every use of the service inside nms

<crweb>

sourcerror: either way you have to connect to something that is running

so now you want to make nms depend on cooler

<crweb>

why?

<crweb>

change it so ir is part of nms, not cooler

you said to change channels

crweb: to use the ir blaster

<crweb>

cooler, UI's etc have to connect nms

crweb: the ir blaster can be used by other apps which don't need to play media

<crweb>

nerochiaro: using cooler?

the UI should not need to be running to execute on scheduled recordings.

<crweb>

sourcerror: excatly

<crweb>

nms is != the gui

crweb: yes

crweb: you can use cooler to use ir blaster to do other things than changing channels

<crweb>

why should i have to icnlude ncooler in my non neuros release?

<crweb>

they are going to make me ship a ui toolkit, so i can channel change?

crweb: make libblaster.so then ?

<crweb>

or build it into the 1 thing everything needs inorder to use media functions

crweb: the way i see it, either separate cooler in smaller libs. or have options to not build parts of cooler

<crweb>

the way i see it is, le nms do its job

<crweb>

leave cooler out of everythin

i agree that gui toolkit and media shall be in different libs

[03:50]

it is still useful to have the scheduler a different process/daemon from nmsd.

[03:50]

<crweb>

why?

[03:51]

they are different concerns.

[03:51]

<crweb>

thats not a reason, and no they aren't

<crweb>

you can't record with out nmsd

that's not a reason

<crweb>

yes it is

either.

sourcerror: when you say "scheduler" you mean the thing that actually starts the recording or the one that manages the database of scheduled recordings ?

manages the database of recordings

sourcerror: then it's even more pointless. since the database is just data. you don't need a running server to manage it

[03:53]

<crweb>

you don't need a daemon for that

[03:53]

<crweb>

its just a sql database

[03:53]

<crweb>

nmsd would read the database, and start the recording

it has to deal with conflicting requests, it should notice if the recording failed for whatever reason and it just calls nms as a service to start and stop according to recording quality etc

<crweb>

nms can tell if it failed

<crweb>

why should nms tell another process that it failed and the other process tell nms what to do?

sourcerror: so it actually engages the recording ,it doesn't just manage data