Log from #neuros IRC channel

MP0: the nms client interface that you mentioned earlier is what you generally need

MP0: what are you trying to implement, if i may ask ?

<MP0>

I wasn't... some other guy asked how to start recording.

<MP0>

without using the GUI

ah, I think he emailed in the list too. "jello", right ?

<MP0>

yes

ok, i'll answer him there then

<MP0>

hopefully he liked the well-commented nms-client.c file

<MP0>

that I pointed him to

MP0: commented, yes. "well", hmmm ;) anyway, it was the right thing to do

thanks

[02:27]

<MP0>

humm, two clients cannot share the same i2c device... oh well, more work.

[02:27]

MP0: sorry to hear that

[03:19]

crweb: can you run a full "build-helper.sh all" build in the background, just to confirm that the main-menu builds normally for you ? for some reason it didn't build for me, very strange

[03:20]

<crweb>

yeah, got to hook up my drive real quick

[03:20]

crweb: no problem if you have to mess up things. just give it a shot if you have it handy

[03:25]

<crweb>

what settings for server/port does your gaim use for gtalk

[03:26]

<crweb>

oh got it

[03:26]

<crweb>

needed ssl enabled

[03:27]

<orion-uk_>

.

[03:27]

<orion-uk_>

[03:35]

wow, that was weird. just got policeman at the door that wanted to check my "office"

[03:37]

<crweb>

check it for what?

[03:37]

check that it existed

[03:38]

<crweb>

ahh.. for taxes?

[03:38]

yep

[03:38]

<crweb>

did it exist?

appearently. the 24" monitor was quite impressive to them

looked like it was the proof they needed

<crweb>

mine doesn't exist...

<crweb>

suite B, is my closet

well, technically mine neither, it's just a room in my home. but that's legal

<crweb>

yeah, usa wants to make sure you have a separate room.

here they are thankfully more lax on that. the boring thing is that i have to go up there tomorrow morning and bring to them a bunch of paperwork.

so i won't be around much

hey nerochiaro and crweb

<crweb>

they'd never come to my house though

[03:41]

daurnimator: hi

[03:41]

<crweb>

daurnimator: hey

[03:42]

crweb: i never expected them either. first time they show up. actually got a bit worried for a minute

[03:42]

<crweb>

i don't think we are required to let them in the house

[03:43]

<crweb>

and any more, they make us go to their office when we get audited

[03:43]

ya, i have to do that tomorrow

[03:44]

regarding inspecting the house, if it's the same thing as the office here they are entitled to enter. they can inspect offices if they want, i think

[03:44]

they actually asked me when i was more comfortable letting them inspect, so i could've just asked them to come another day and shredded all compromising evidence, if i had any

[03:51]

<crweb>

JohnDesigner: ping

[03:51]

<crweb>

if that man has any sense. he's sleeping..

[03:55]

crweb: are you following that ENTER/# key thing yourself ? I don't really understand that last email where he says "it's unsensable"

[03:55]

<crweb>

he's combining 2 topics

[03:55]

<crweb>

the ability to send the key back, and linking # and ENTER together

[03:57]

<crweb>

virtual keyboards don't send the key back for a reason, which is what I just explained. This is a john call though and should be implemented in NLineEdit, not the VKB (since other users/apps might not want this "enter key")

[03:57]

<crweb>

i failed to metion the case where, there are some widgets where sending "Enter" again would cause an infinite loop..

[03:58]

<crweb>

a LineEdit with a button buddy. for example

[03:58]

<crweb>

button opens keyboard. "Enter" in line edit sends input to button, which opens the keyboard

[04:01]

be sure he doesn't get tripped up by that issue, please

[04:01]

and if you can clear up the mixing of concepts, that would help a lot too

[04:02]

<crweb>

I told him we'd implement that later and to go ahead without it already.

[04:02]

ok, i see that hunt is in agreement too

[04:02]

<crweb>

I'll wait for john to wake up and then we can continue

[04:07]

crweb: it's ok

[04:20]

crweb: did you happen to try that build-helper.sh all ?

<crweb>

it is going at the moment

<crweb>

agh. faild at no libz

<crweb>

how odd

for me it suceeds, but it doesn't build main-menu at all

[04:22]

and if i build it manually, it gives me a lot of mediakey-related errors while running

[04:22]

and yet, nick did not commit his patch yet

[04:24]

crweb: did you run a "build-helper up" before building ?

[04:24]

<crweb>

i did yes

[04:25]

<crweb>

i'm missing something with libz on my system..

[04:25]

<crweb>

someone posted notes in the guide about it i think

[04:25]

yes, i think you just have to install it in your distribution

[04:28]

<crweb>

zlib1g

[04:29]

<crweb>

err, ?

[04:29]

<crweb>

i've never had to install that before... i don't get that..

[04:29]

i've never had to install it at all

maybe you uninstalled something that auto-uninstalled that lib too

dunno

anyway, check what happens now that you can run a full build again

<crweb>

now i don't follow why # must = Enter

<crweb>

that makes no sense.. no user things.. hmm I'm going to press #

<crweb>

err, thinks

i totally have no idea

i asked if it was from John's design

but i don't think he answered

<crweb>

i think he got busy with my replies and missed your question

<crweb>

who presses #....

[04:32]

well, if it's not in John's design, it has to be removed (and i'm thinking it has to be removed anyway)

[04:32]

<crweb>

never woui think we all agree with my "Enter" doesn't mean auto activate action

[04:32]

<crweb>

ahah. remove the "never wou"

[04:34]

well, softkeyboard has and ENTER virtual button. if you press that, it close softkey and input text into widget. what is the issue ?

[04:34]

<crweb>

he wants that "Enter" to auto activate an action. Not _just_ put text in the widget

[04:34]

<crweb>

and instead of having to press "Enter" on VKB. you can just press #

[04:35]

<crweb>

but who knows that.

[04:35]

last time i checked # on virtual keyb was to switch between modes (upcase, lowcase, symbols)

[04:36]

as for auto-activate action, i'm not sure what you mean. you put text into widget, and that's it. what action ?

[04:36]

<crweb>

if text was Youtube Search, pressing VKB "Enter" would send the enter key to the LineEdit to activate the search

[04:37]

<crweb>

instead of only placing text in the line edit

[04:37]

<crweb>

but my beef with that is. what if i typo. I have to wait for long search to finish to edit?

[04:37]

<crweb>

that sort of logic pisses me off bad.

[04:38]

<crweb>

could just be me though..

[04:38]

it just doesn't make sense

<crweb>

yeah, thats more of a Virtual Input/Action handler. not a VKB

softkeyb closing and inputting text should be a different thing than any other action from the widget that has that softkey widget

i mean, you have usually to press an "OK" button, or "Search" button on the search form

<crweb>

the application should listen to the signal newText(), and "activate on new text"

<crweb>

not VKB take action, or Send action key by default

[04:40]

<crweb>

I am not going to allow auto action in NLineEdit/VKB. I just proved it not needed. Plus, there are more edits that don't need auto action than edits that do

[04:40]

whatever, it depends on the app design, but it's unrelated to softkeyb. the fact is: softkeyboard should only care about inputting text into widget.

<crweb>

agree

application takes care of actions

softkey doesnt' even know about what app it is being called from

how can it trigger actions in the app ?

<crweb>

he wants to send the Enter key to the NListEdit on VKB close

<crweb>

and NListEdit would take action

just make a specific signal that say "i finished inputting text". then each app can take the action it wants

<crweb>

already exists

or simply if it's QT standard just attach to the text changed signal

i bet there's one

<crweb>

thats what you do yes

<crweb>

there is, it was newText() i think

ok, please make sure he does that and understand is

er, it

[04:43]

<crweb>

so, that pretty much closes the Issue

[04:43]

<crweb>

# handling would be in the VKB

[04:43]

<crweb>

and is a different topic

[04:44]

yes

[04:44]

case closed, if you can keep him on the resolution and hopefully make him understand it

[04:46]

now, about that build... this is freaking me out

[04:50]

<crweb>

nerochiaro: i built fine btw

[04:51]

<crweb>

completely ok

[04:52]

does it run ok too ?

crweb: is main-menu application present '

?

<crweb>

wtf

neuros-bsp/rootfs/fs/usr/local/bin/main-menu

i don't have this

<crweb>

how did that happen

<crweb>

build should fail..

scary eh ?

<crweb>

it didn't build at all

if you build it manually "make -C linux-r3-main-app all install" it won't create it either. you have to "make -C linux-r3-main-app/main-menu all install" to make it work

<crweb>

i went and typed make and it built fine

ya

<crweb>

so.. who screwed up linux-r3-main-app/Makefile ?

i checked it, but i don't see screweups

let me check the svn log

[04:57]

<crweb>

its got to be jimmy

[04:57]

last commits are his

[04:58]

<crweb>

4353, should that be a \ before done?

[04:58]

<crweb>

nvrmind, guess so.

[04:58]

<crweb>

it all looks like copy and paste to me...

[04:59]

anyway, he just added qm_install, didn't touch install itself

[04:59]

<crweb>

its not even being compiled.. must be pre-qm_install

[04:59]

ah, i get it

4393 jimmy lrelease translations/*.ts || \

3594 mgao make all && cd .. || exit 1 ;\

that means

that if lrelease fails, it won't get built

but it won't fail the build

so, two questions:

<crweb>

linux-r3-main-app/Makefile should error out though

why ?

the || will prevent anthing else from executing

and will return "ok" return code

<crweb>

right. if it errors, the make file isn't suppose to continue building the other stuff

but it dosn't

error

he should just use && in place of ||

<crweb>

oh i see

or even better

lrelease || exit 1

and then "make all" and so one

let me see if this fixes it

[05:03]

<crweb>

crap

[05:03]

<crweb>

make all, in linux-r3-main-app just worked for me

[05:03]

<crweb>

didn't work from build-helper though..

[05:04]

<crweb>

checking make install now.

[05:06]

i patched it: it correctly fails for me now telling me "lrelease error: Cannot open file 'translations/*.ts': No such file or directory"

[05:06]

this is at least the correct error message

[05:07]

<crweb>

make install is failing for me. because main-menu is not completely building.

[05:07]

<crweb>

after you make all in main-app, go into main-menu and type make

[05:07]

that is not the real issue, i think

<crweb>

the install isn't finding the files because they didn't exist.

the real issue is that it should quit immediately when lrelease fail

which files ?

<crweb>

main-menu/build/*

<crweb>

there is a 4 step process in main-menu/*

<crweb>

nqmake -recursive

<crweb>

make

<crweb>

nqmake -recursive

<crweb>

make install

<crweb>

make install won't install files that didn't exist the last time qmake was run

of course they didn't exist, they were never built. and they were not built because lrelease failed failed on main menu, but instead of failing all build, it just skipped building main-menu instead

let me send you a patch for this

<crweb>

lrelease should be run after make

ah

<crweb>

not during the compile..

that is another error then

so first make, then lrelease ?

<crweb>

yeah

and if any fail, stop the build

let me try that

[05:11]

<crweb>

that way they are separate. and main-menu will still build even if language fails

[05:11]

i would stop the build if lrelease fails, though

[05:11]

<crweb>

right, but as we see. which caused the build to fail?

[05:12]

<crweb>

much harder ti diag when they are both missing

[05:12]

<crweb>

also, the language install and application build are very separate things. applications can build and install without language packs

[05:12]

so, what do you suggest ?

<crweb>

all language generation done AFTER application make and make install ?

but do we still fail the build if they fail ?

<crweb>

or at least let the application finish building

<crweb>

yes

[05:14]

<crweb>

if language fails to build and script stops

[05:14]

so what's the need to do it after all apps ? we can do it after each app and stop build immediately on that app

[05:14]

<crweb>

i still want main-menu/build/* to exist have have built app in it

<crweb>

true, if it all fails it won't matter cause you can just read where it failed

<crweb>

i'm wondering if we should just let it continue without language packs

i think we should not

<crweb>

ok

<crweb>

kill it then, and just make it fail where ever you can

[05:16]

crweb: sendign you patch in a minute to double check

[05:16]

<crweb>

ok

[05:16]

<crweb>

ahh, i was looking in qm_install

[05:17]

<crweb>

i see the lrelease stuff now

[05:17]

<crweb>

patch qm_install also to exit 1

[05:17]

yes

<crweb>

no here is another issue:

crweb: wait

<crweb>

oh nvrmind

<crweb>

i miss read

[05:19]

<crweb>

hahaha

[05:19]

i created the patch, which cause it to run (for each app) first "make all", then lrelease translations/*.ts (and fail the whole build if any fails). but it is failing on me

<crweb>

i found it.

on main-menu's lrelease

<crweb>

you can't *.ts in qmake file

<crweb>

the .ts files should be added to the ui.pro individually so they can be generated

<crweb>

you can't generate a *.ts file

[05:21]

what is that is supposed to generate the *.ts files ?

[05:21]

<crweb>

qmake does

<crweb>

kind of. lrelease reads the .pro file for available translations

<crweb>

you don't tell it "*.ts" if i remember correctly

<crweb>

i will continue work on this

<crweb>

he left out some key pieces

ok. do you want the patch i made so far ?

<crweb>

you just added exits ?

more or less, reordered these || and && that were not right

<crweb>

yes then

<crweb>

he never runs lupdate to generate the ts files.

sent now

please notice the notes in the email, too

<crweb>

we are suppose to commit the ts files to the tree. he must have missed this

<crweb>

step 1: commit clean .ts files generated from source

<crweb>

step 2: email .ts file to translator

<crweb>

step 3: recieve translated .ts file from translator

<crweb>

step 4: commit translator .ts file to tree

<crweb>

step 5: run lupdate

<crweb>

step 6: send translator updated .ts

<crweb>

step 7: recieve translated .ts file

<crweb>

step 8: commit translated .ts file

<crweb>

Repeat 5-8

let's make this clear then

and fix this mess by failing when they are not there

so he's forced to commit them anyway

<crweb>

well, after first commit, they will never stop existing

<crweb>

he just didn't finish

<crweb>

but, yes. fail away

[05:29]

it's good to fail, in case they get deleted for some reason

[05:40]

crweb: still there ?

[05:41]

<crweb>

yes

[05:55]

crweb: pong

[05:55]

<crweb>

JohnDesigner: please hold. I'm not in the mood for your shananigans

[05:55]

?

[05:56]

<crweb>

yeah i dunno. I'm still not taking JohnDesigner input..

[05:56]

ah, I see.

[05:56]

<crweb>

i'll think about it in a few min when done meeting with pablo

[05:57]

still acting like the gal that got felt up at the movie theater I see

[05:57]

:)

[05:58]

so , #= enter on the softkeyboard. Is that the problem?

<crweb>

you just pick that up from IRC, or did someone beat me to the email?

IRC

<crweb>

not to much a problem

<crweb>

just not expected behavior

<crweb>

I'd never have figured out # was "Enter"

well...

what do you sugges.

t

?

<crweb>

just leave it VKB "enter" box

[06:00]

<crweb>

this is really a minor issue. It would be easy to add to VKB either way. There is much larger issue

[06:00]

no "enter" shortcut? You

would rather navigate to the |enter| key each time?

<crweb>

i do anyway

<crweb>

a.) never knew # was (enter). b.) I'll never remember that..

what is the big deal with #?

<crweb>

nothing, forget about it. We can do it.

no, I want to hear it

[06:02]

<crweb>

it was tied into a larger issue from xiamen we had to decode

[06:02]

<crweb>

once we figured out real issue, that one became nothing..

[06:02]

ah, i see.

Hmm, I dont have a remote in front of me, is the # key on the right or left?

I cant remember

<crweb>

pfft

pfft?

<crweb>

like i would know

where is it on your cellphone ?

do you have a remote in front of you?

i have one at my side

does it count ?

<crweb>

its on the right

<crweb>

funny there are 2 * keys

<crweb>

oh wait.. thats an xim

* crweb throws john a cheap shot

[06:05]

* JohnDesigner is leaving

[06:06]

<crweb>

JohnDesigner: my main issue with # being enter is that, # clearly says Clear on the button

[06:06]

<crweb>

and I'd expect it to clear. boy would i be pissed if it Entered and auto activated my search

[06:07]

its a shortcut

[06:07]

<crweb>

leading to, I'd never would have thought of hitting the "Clear" button to (Enter)

<crweb>

anyway, where ever you want to put it. whatever.

that is why there are words on the screen "press # to enter".

<crweb>

Main issue:, auto activation

no, its not a perfect situation.

[06:09]

no, I would not do that if designing the system and remote from scratch.

[06:09]

<crweb>

for this conversation VKB stands for Virtual Keyboard. Activation means that a task is started

[06:09]

mmhmm

[06:10]

auto activation?

[06:11]

<crweb>

Example: Youtube Search. User brings up VKB, enters text. Presses Enter(#), VKB closes and activates search

[06:11]

I don't see what the big deal is with the shortcut.

<crweb>

forget the shortcut

<crweb>

we'll do that

okay, but it was in your last questions

<crweb>

that was just a statement of events