[02:12]	<nerochiaro>	crweb: you around ?
[02:14]	<crweb>	nerochiaro: yes
[02:14]	<nerochiaro>	crweb: you think it's a good idea to make our spinedits emit a signal when a value is input that is out of bounds ?
[02:15]	<crweb>	no
[02:15]	<crweb>	there is no bound
[02:15]	<nerochiaro>	maxvalue
[02:15]	<crweb>	the spinedits can contain any data
[02:15]	<nerochiaro>	no, check setRange, setMinimum, setMaximum
[02:15]	<nerochiaro>	numeric spinedits
[02:15]	<crweb>	its not that kind of spin edit
[02:15]	<nerochiaro>	it is
[02:16]	<crweb>	nhorzspinbox ?
[02:16]	<nerochiaro>	aye
[02:16]	<crweb>	it is not
[02:16]	<crweb>	it is a model/view
[02:16]	<crweb>	and can contain any data
[02:16]	<crweb>	including Text strings for "English","French","Polish", etc
[02:16]	<crweb>	how do you maxValue that?
[02:18]	<crweb>	he should subclass QSpinBox, and draw it horizontally
[02:18]	<nerochiaro>	enum DisplayRole -> DR_NUMBER as opposed to DR_STRING. then it will accept numeric input from number keypad and it will have max and min values
[02:18]	<nerochiaro>	check the class
[02:18]	<nerochiaro>	NHorizontalSpinBox
[02:18]	<crweb>	well, then its incorrect
[02:19]	<nerochiaro>	well, then it may have to be fixed by splitting the class. but as it stands now, the out of bounds signal kind of make sense to me
[02:19]	<crweb>	a numeric spinbox is very different from a variable data spinbox
[02:19]	<nerochiaro>	i agree there. right now it's handled by that enum, even though they should be separate widgets
[02:20]	<crweb>	if that feature is already implemented then you don't have to emit a signal
[02:20]	<crweb>	but you can
[02:20]	<crweb>	the spinbox just can't move higher than max value
[02:20]	<crweb>	and if it does, make it max value
[02:20]	<nerochiaro>	i know, it's already that way
[02:20]	<crweb>	I think its obvious to a user that is the maximum value
[02:21]	<crweb>	i mean, they cannot be that stupid..
[02:21]	<nerochiaro>	ok
[02:22]	<crweb>	jason's got me all worked up on how poorly we treat customer because we treat them as stupid on simple things and drop them on advanced things.
[02:22]	<crweb>	;O
[02:22]	<crweb>	btw, I'm thinging that nwm isn't the leak
[02:22]	<dav7>	hmm
[02:22]	<crweb>	for 4349
[02:22]	* dav7  is a UI developer
[02:22]	* dav7  considers how to make it more obvious that a spinbox can't go higher
[02:23]	<crweb>	dav7: don't let the user type any higher
[02:23]	<dav7>	heh
[02:23]	<dav7>	I KNOW
[02:23]	<crweb>	nerochiaro: actually you can do that
[02:23]	<crweb>	you can create a validator to apply on the input that wouldn't allow typing any higher than that value
[02:24]	<crweb>	but I think we have a better solution
[02:24]	<dav7>	instead of showing just a single value, show 1 and two halves of a value: half of the previous one (at the bottom) and half of the next one (at the top)
[02:24]	<crweb>	I'd get pissed cause it wouldn't type
[02:24]	<dav7>	when we try to go a little higher, do a little bounce-back animation, indicating we've hit a "wall"
[02:24]	<crweb>	nerochiaro: or show the max and min value
[02:24]	<dav7>	does that work?
[02:24]	<dav7>	does/would
[02:25]	<crweb>	dav7: interesting
[02:25]	<dav7>	thanks :D
[02:25]	<nerochiaro>	dav7: i actually made it just go back to the max allowed value for now. if max is 55 and user type 99, it will input 55 automatically.
[02:25]	<dav7>	right.
[02:26]	<crweb>	nerochiaro: all that error about cannot create pipe just means is that something has opened all the descripters and nwm is the one that hit the wall.
[02:26]	<dav7>	add a little animation to that where it vibrates a little and "inverts" to red (like the screen does if you die in a 3D game eg)
[02:26]	<dav7>	^ nerochiaro
[02:26]	<dav7>	vibration/blur/wobble effect, etc
[02:26]	<nerochiaro>	dav7: and displayed on top of the spinbox "you can't input a value higher than X" for good measure. the anim is nice, will consider it for future
[02:26]	<dav7>	right. cool :P
[02:27]	<nerochiaro>	crweb: who is hogging these file descriptors then ?
[02:27]	<nerochiaro>	crweb: it has to be either scheduler or nwm, i don't see any other suspects tehre
[02:28]	<crweb>	I think its nmsd
[02:28]	<crweb>	I am going to have Sean make a regression test. Run same test on torfu
[02:28]	<crweb>	I cannot make this bug occur so far
[02:28]	<crweb>	and there are 8192 file descriptors available
[02:29]	<dav7>	oh btw, with my other idea, when you've hit the max value, there isn't another value "above" the current one, to indicate you can't go in that direction any more. If the scrolley/half values idea and the flicker-to-red ideas were merged, if the user types a wrong value have it "scroll" down (quickly) to the highest available value.
[02:29]	<dav7>	maybe.
[02:29]	<crweb>	nwm doesn't open enough QProcesses to hit the limit, and vrecorder closes every run
[02:29]	<crweb>	the only thing left open is nmsd
[02:31]	<crweb>	without lsof i think it maybe hard to tell
[02:35]	<nerochiaro>	dav7: hard to do with our UI i think, but nice
[02:35]	<nerochiaro>	crweb: doesn't bbox have lsof ?
[02:35]	<crweb>	nope
[02:35]	<nerochiaro>	maybe it's enough to compile it ?
[02:36]	<nerochiaro>	it's a small util
[02:36]	<crweb>	yeah, its not written very compilable
[02:36]	<crweb>	trying now though
[02:36]	<nerochiaro>	in any case, if it's nmsd, it's easy to find out. write a test app that start and stop recording a lot of times in a row and see what happens
[02:36]	<nerochiaro>	it may well be some file descriptor we forget to close
[02:37]	<nerochiaro>	maybe some of the mp4 scratch files, or something
[02:37]	<crweb>	thats what i was thinking yeah
[02:37]	<nerochiaro>	try asf recording too and see if happens or not
[02:37]	<crweb>	unfortunately, 4349 is 2 bugs
[02:38]	<crweb>	the file descriptor only happened in chains test. the real bug it did not happen
[02:41]	<nerochiaro>	what would be the "real bug" ?
[02:41]	<crweb>	4349
[02:43]	<nerochiaro>	it seems to me there's only the bug about new recordings can't be started because of file handler leak
[02:43]	<crweb>	thats not true
[02:43]	<crweb>	as i just said. 4349 has no info about file descriptors
[02:43]	<crweb>	until Chain posted a comment
[02:43]	<crweb>	file descriptors was not in the terminal output
[02:44]	<nerochiaro>	i couldn't check that log that she posted on a share instead of attaching it
[02:45]	<crweb>	i had her email them to me
[02:45]	<crweb>	chain found something else
[02:45]	<crweb>	somewhere
[02:45]	<crweb>	we don't know what he did, or what he was doing
[02:45]	<crweb>	maybe *he* was holding all the descriptors
[02:46]	<crweb>	4003 was the result of the same test on TORFU
[02:47]	<crweb>	that one is just strange
[02:49]	<nerochiaro>	beside the MS card freeze up you mean ?
[02:49]	<crweb>	?
[02:49]	<nerochiaro>	oh, nevermind
[02:50]	<nerochiaro>	might it be it can't restart because scheduler is trying to record immediately ?
[02:50]	<nerochiaro>	hmm, no
[02:50]	<crweb>	I don't even see the crash though
[02:50]	<crweb>	Finished MP4 Finalization, returned 0
[02:51]	<crweb>	very bottom of the first term log
[02:51]	<nerochiaro>	that seems kosher to me
[02:51]	<nerochiaro>	recording completed ok
[02:51]	<crweb>	then she says it crashed
[02:51]	<crweb>	the ps -A shows everything running though
[02:52]	<nerochiaro>	so it actually hang
[02:52]	<nerochiaro>	but why a reboot doesn't fix, beats me
[02:52]	<crweb>	well, i see bad nand
[02:52]	<crweb>	and lots of corrupt symbols
[02:52]	<crweb>	could be her PC though
[02:53]	<crweb>	i think she ran /mnt/OSD out of space maybe
[02:53]	<crweb>	I don't know what that error looks like on torfu though
[02:57]	<nerochiaro>	the bogus symbols are just an issue with terminal
[02:57]	<nerochiaro>	the bad nand... well, we've been told not to worry
[03:06]	<crweb>	recheck_schedule no item........................
[03:06]	<crweb>	There is the issue
[03:07]	<crweb>	Scheduler didn't find a new item to schedule in Comment #7
[03:07]	<crweb>	even though there were 30 left
[03:07]	<crweb>	Or, recorder didn't find the next item
[03:08]	<crweb>	nope, scheduler
[03:08]	<crweb>	that is confirmed
[03:09]	<nerochiaro>	didn't find or couldn't execute ? and TORFU or AZ bug ?
[03:09]	<crweb>	AZ, didn't find
[03:09]	<crweb>	4349
[03:09]	<crweb>	Log of comment #7
[03:10]	<crweb>	This coencides with JoeBorn's complaints
[03:10]	<crweb>	Either they created them wrong, or scheduler stopped looking
[03:11]	<nerochiaro>	or didn't really extract them
[03:12]	<nerochiaro>	i mean, it looked for new items with a query and didn't find any
[03:12]	<nerochiaro>	when in fact there were more
[03:13]	<crweb>	maybe
[03:13]	<crweb>	maybe there were more i mean
[03:13]	<crweb>	maybe the application used to add schedules ignored conflicts
[03:13]	<nerochiaro>	i think it used the regular API to add them. who made that app ?
[03:14]	<crweb>	no idea, and they don't
[03:14]	<crweb>	erry, they don't, and no idea
[03:15]	<nerochiaro>	why don't you just ask that in the bug '
[03:15]	<nerochiaro>	?
[03:15]	<crweb>	workin on it
[03:15]	<nerochiaro>	oh, ok
[03:15]	<crweb>	you respond much faster than them ;)
[03:17]	<nerochiaro>	but i don't know the answer
[03:18]	<crweb>	If i didn't flood IRC, i just wouldn't feel like i was working.
[03:19]	<nerochiaro>	nah, don't worry
[03:23]	<crweb>	nerochiaro: do you remember how to get to the internal addresses at xiamen?
[03:24]	<nerochiaro>	crweb: i do, but sometimes peter forgets to bring up the NFS mount that allow access to that
[03:24]	<nerochiaro>	crweb: hold a second
[03:25]	<nerochiaro>	59.60.10.118:/home/neuros/mytest-files
[03:25]	<nerochiaro>	i can mount that with sshfs on the good days
[03:25]	<crweb>	hmm.. ok
[03:25]	<nerochiaro>	if it doesn't work, tell me and i'll pester Peter
[03:26]	<nerochiaro>	doesn't work, bugging peter now
[03:26]	<crweb>	good luck
[03:26]	<crweb>	been begging for him all morning
[03:26]	<nerochiaro>	i just got him a few minutes ago to fix the ML
[03:27]	<crweb>	i think he took the opertunity to skadaddle
[03:27]	<crweb>	opportunity even
[03:28]	<nerochiaro>	corlin said he's on vacation this afternoon
[03:28]	<nerochiaro>	he just left
[03:28]	<crweb>	what did i say to corlin earlier
[03:29]	<crweb>	You know, its pretty neat that Neuros always has someone working 24 hr a day
[03:29]	<crweb>	but even with someone on duty at all hours, the person you need is never on duty.
[03:29]	<nerochiaro>	blame it on globalization ;)
[03:30]	<crweb>	when you right click on a tar/bz/gz can you select "Extract"
[03:31]	<nerochiaro>	where ?
[03:31]	<crweb>	on desktop (Debian)
[03:32]	<crweb>	or in file browser (Debian)
[03:34]	<nerochiaro>	crweb: on a .zip it give me "extract here"
[03:35]	<crweb>	hah
[03:35]	<crweb>	it was rar, and I didn't have unrar installed
[03:35]	<nerochiaro>	that's debian default yes
[03:35]	<crweb>	i just don't see "extract here" as often as I did in buntu
[03:37]	<crweb>	well, i can see why it kind of fails (the schedule stuff)
[03:37]	<crweb>	it manually adds the schedules to the db
[03:37]	<crweb>	and for each schedule it sends another QCop saying to scheduler "recheck"
[03:38]	<crweb>	eventually, scheduler shuts down cause recording is started
[03:38]	<crweb>	but we send recheck
[03:38]	<crweb>	we should never send recheck when mass filling the db
[03:38]	<crweb>	restart the unit when you are done....
[03:39]	<crweb>	so everything is just like a user would have it
[03:39]	<nerochiaro>	so the test was done with an invalid program that didn't use the API to fill the DB ?
[03:39]	<nerochiaro>	er, to add schedules ?
[03:39]	<crweb>	yeah
[03:40]	<crweb>	whoever wrote it injected the schedules into the db
[03:40]	<crweb>	its not the testers fault. They'd have no way to know
[03:41]	<crweb>	I bet you could never guess who wrote this
[03:42]	<nerochiaro>	crweb: i know
[03:43]	<nerochiaro>	crweb: then let's have someone try again with a proper program
[03:46]	<nerochiaro>	crweb: is it kosher to add nfs mounts to the fstab ?
[03:46]	<crweb>	yes
[03:46]	<nerochiaro>	sane options i should use, like nolock ?
[03:47]	<crweb>	i use the same ones used on osd
[03:47]	<crweb>	memory fades what they are now
[03:48]	<crweb>	you should also add a timeout
[03:48]	<nerochiaro>	np, i have them
[03:48]	<crweb>	man mount
[03:50]	<nerochiaro>	crweb: it's already there, but how to make it remount when the share go down ?
[03:51]	<crweb>	automount
[03:51]	<crweb>	autofs
[03:51]	<crweb>	Or, don't umount it
[03:51]	<crweb>	when it comes back up it'll just start working again
[03:51]	<nerochiaro>	in any case, it should work now, if you try ti
[03:51]	<nerochiaro>	it
[03:51]	<nerochiaro>	let me know if it has the stuff you need
[03:53]	<crweb>	aparently i'm not part of the neuros group
[03:53]	<crweb>	which pretty much sux since I can't cd into anything
[03:53]	<crweb>	oh, i'm not in postdrop
[03:53]	<crweb>	owner of mytest-files
[03:54]	<nerochiaro>	you can't cd ?
[03:54]	<crweb>	i can't access mytest-files no
[03:54]	<nerochiaro>	how do you mount it ?
[03:54]	<crweb>	owner steven.zhang:postdrop
[03:54]	<nerochiaro>	ah, you just ssh in the box
[03:55]	<crweb>	yeah.. i'mnot really into the whole virt fs thing
[03:55]	<nerochiaro>	i can add you to groups
[03:58]	<nerochiaro>	try now
[03:58]	<nerochiaro>	you're in postdrop too
[04:04]	<crweb>	"you got the power"
[04:04]	<nerochiaro>	:D
[04:05]	<crweb>	no there is nothing in there that i need
[04:05]	<crweb>	i need neurosshare\osd-3\osd10
[04:07]	<crweb>	nerochiaro: should we hold off on writing a new tool till next week when we have a dev that can do so?
[04:07]	<nerochiaro>	crweb: i wouldn't use him for this stuff
[04:07]	<crweb>	well, he's probably going to change a few things
[04:08]	<nerochiaro>	well, ok, can be part of the scheduler fixing job
[04:08]	<crweb>	hopefully the bug would be come invalid automagically
[04:10]	<nerochiaro>	crweb: hopefully
[04:12]	<nerochiaro>	59.60.10.118:/home/neuros/internal-share
[04:12]	<nerochiaro>	give it a shot
[04:18]	<nerochiaro>	crweb: and about jwu's lack of commit msg, he's right. the msg is there in the Subject field
[04:18]	<crweb>	i still can't see it
[04:18]	<crweb>	why is it different than the others?
[04:18]	<nerochiaro>	Subject: [PATCH] copy upk tool sources from build-tools
[04:19]	<crweb>	ah crap. I read that as email subject
[04:19]	<nerochiaro>	me too :)
[04:20]	<crweb>	my first public post...
[04:20]	<crweb>	awesome
[04:20]	* crweb  is retarded folks
[04:21]	<nerochiaro>	np
[04:23]	<crweb>	well, do you have any pressing tasks?
[04:24]	<crweb>	Corlin just gave my my schedule
[04:25]	<crweb>	and then told me. That is all you are doing next week. I assign you no more.
[04:26]	<crweb>	i dunno what I'm going to do with the extra 8 hours of daytime next week...
[04:27]	<crweb>	err, extra 8 hours a day
[04:27]	<crweb>	8 work, 16 off.. by wed. i'll be insane ;)
[04:29]	<crweb>	nerochiaro: and if corlin didn't mention it to you. AZ4 and AZ5 are leaning toward refactoring and enhancements.
[04:30]	<nerochiaro>	that would be nice
[04:31]	<nerochiaro>	as for tasks, are you still looking at OOM stuff ?
[04:31]	<crweb>	yep
[04:31]	<nerochiaro>	can you also look at that QT issue that it can't update correctly the dir view when we are browsing a NAND dir ?
[04:32]	<crweb>	language, oom, nwm leak, scheduler refactor, decrease memory usage, osd2 rootfs, osd2 qt port, osd2 toolchain
[04:32]	<crweb>	I can try yeah
[04:32]	<crweb>	without nand, its about damn impossible to duplicate.
[04:33]	<nerochiaro>	crweb: scheduler refactor shouldn't be on you
[04:33]	<crweb>	its not
[04:33]	<nerochiaro>	you put in the list one moment ago
[04:33]	<crweb>	but someone has to bring him up to speed etc
[04:33]	<nerochiaro>	ok
[04:34]	<nerochiaro>	and what's up on oom, what's the latest news ?
[04:34]	<crweb>	can't duplicate
[04:34]	<crweb>	i'm well past 70 hours on 6 tests
[04:34]	<crweb>	I'm actually setting up a package so that xiamen can automate the dumping of information i actually need
[04:35]	<nerochiaro>	you know you never will duplicate it yourself, it's a random bug and Murphy's law will never allow that to happen to you when you need it
[04:35]	<crweb>	i read though that file descripters can cause OOM
[04:35]	<nerochiaro>	good idea on the package
[04:35]	<crweb>	which is what really has be curious
[04:35]	<nerochiaro>	how can they ?
[04:36]	<crweb>	technically, its a memleak
[04:36]	<crweb>	the randomness regardless of setup has nmsd screaming all over it
[04:37]	<nerochiaro>	didn't we say it doesn't happen with just vrecorder -qws and nothing else ?
[04:37]	<crweb>	sure, we said that
[04:38]	<nerochiaro>	but you're not sure it's true ?
[04:38]	<crweb>	i think assuming that is a solid base is a mistake
[04:38]	<nerochiaro>	why ?
[04:38]	<crweb>	technition intuition.
[04:39]	<crweb>	technician even
[04:39]	<nerochiaro>	hmm
[04:39]	<crweb>	6 tests with a bug this random.. doesn't prove vrecorder won't OOM
[04:40]	<nerochiaro>	right, especially with your own results
[04:40]	<crweb>	I've done 6 tests with EVERYTHING running
[04:40]	<crweb>	xmms, djmount, nwm, main-menu, scheduler, vrecorder,
[04:41]	<crweb>	I'm going to switch to testing from my flash though
[04:41]	<crweb>	1/2 tests flash, 1/2 tests nfs
[04:42]	<crweb>	i've got to recreate whatever it is they have in xiamen that I don't
[04:42]	<crweb>	like.. a CF driver
[04:43]	<nerochiaro>	a CF driver ?
[04:43]	<crweb>	if you run in flash, scratch area is on CF
[04:43]	<crweb>	right?
[04:44]	<crweb>	or is it still on the media being written to
[04:44]	<nerochiaro>	CF or NAND
[04:45]	<nerochiaro>	/media/ext/.osd-extended/scratch/
[04:46]	<crweb>	so xiamen tests are using media/nand, and i'm using nfs
[04:46]	<nerochiaro>	you can still run off of nfs for the rootfs and mount the CF as normal instead of mounting ext area off of nfs, btw
[04:46]	<crweb>	i know it sounds weird, but duplicating them is going to have to be a must
[04:47]	<crweb>	anything different is a variable that could be missed
[04:47]	<nerochiaro>	duplicating the bugs ? or the setup ì'
[04:47]	<nerochiaro>	^
[04:47]	<nerochiaro>	?
[04:47]	<crweb>	the setup
[04:48]	<nerochiaro>	well, the nfs-mounted rootfs i don't see how it can be a problem, as long as scratch is CF-mounted
[04:48]	<nerochiaro>	and besides, it may be that chain in his tests was nfs-based
[04:48]	<nerochiaro>	i didn't ask
[04:48]	<crweb>	I can't either but when all else fails duplicate and repeat.
[04:49]	<nerochiaro>	do you know what he was using then ?
[04:49]	<nerochiaro>	NAND or CF or nfs ?
[04:49]	<crweb>	no, I know testing was using NAND
[04:49]	<nerochiaro>	chain's testing ?
[04:49]	<crweb>	no team
[04:49]	<nerochiaro>	ah, ok. and you don't have nand units ?
[04:50]	<crweb>	sorry, there were 4 nand and 5 CF
[04:50]	<crweb>	they can't tell me which failed more or didn't fail
[04:50]	<crweb>	the only know some failed
[04:51]	<nerochiaro>	i see. and you don't have NAND unit there, right ?
[04:51]	<nerochiaro>	can't you get Chicago to send you one ?
[04:51]	<crweb>	right
[04:51]	<crweb>	they don't exist yet
[04:51]	<nerochiaro>	xiamen has some
[04:51]	<crweb>	I've been barking up that tree for 3 months
[04:52]	<nerochiaro>	that sounds wrong. added to TODO to find out why the hell you don't have one yet. i'll get back to you
[04:52]	<crweb>	add the same item to your TODO for yourself
[04:52]	<nerochiaro>	to get one ?
[04:52]	<crweb>	yeah
[04:53]	<nerochiaro>	sure
[04:53]	<crweb>	I'm going to be a little upset if I have to go buy one to get one ;)
[04:53]	<nerochiaro>	on an unrelated note, do you have any idea about this ? forums.neurostechnology.com
[04:55]	<crweb>	nerochiaro: i do not.
[04:55]	<crweb>	nerochiaro: I haven't even heard of that before
[04:56]	<nerochiaro>	me neither. the only thing that look suspicious there is the WD USB disk. it seem people are having issues with these, even in ML
[04:56]	<crweb>	yeah but thats all I own
[04:57]	<nerochiaro>	maybe you have older models, i dunno
[04:57]	<crweb>	it is possible though that WD did something
[04:57]	<nerochiaro>	or access then via USB enclosures ?
[04:57]	<nerochiaro>	them
[04:57]	<crweb>	they have become shady lately with some of their contracts with media companies
[04:57]	<crweb>	yeah i have 6 MyBook drives
[04:58]	<crweb>	3 office, 3 home
[04:58]	<crweb>	(those are models)
[04:59]	<crweb>	I would say TJS should try to playback from different media
[05:00]	<nerochiaro>	he is going to try that. see last post in thread
[05:01]	<crweb>	i mTJS or Caterpillar?
[05:04]	<crweb>	nerochiaro: going out to breakfast
[05:12]	<nerochiaro>	crweb: later
[07:04]	<nerochiaro>	crweb: you around ?
[07:05]	<vmarks>	hi
[07:07]	<nerochiaro>	hi
[07:09]	<daurnimator>	www.jeremygaddis.com
[07:10]	<greyback>	daurnimator: lol
[09:12]	<jpsaman>	hu
[09:14]	<nerochiaro>	jpsaman: he
[09:19]	<jpsaman>	hi
[09:58]	<jpsaman>	nerochiaro: meeting was at 17 CET on #neuros ?
[09:58]	<nerochiaro>	jpsaman: yes
[09:58]