Camelia, the Perl 6 bug

IRC log for #motoactv, 2012-04-27

| Channels | #motoactv index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:03 CEnnis91|Mobile joined #motoactv
01:11 Quietlyawesome94 joined #motoactv
01:30 [mbm] CEnnis91: http://pastebin.com/SMwQqstq
01:31 [mbm] not sure why that file is 257M
01:32 [mbm] weird; remade the zip file and now it's 162M
01:45 Quietlyawesome94 joined #motoactv
02:18 CEnnis91|Mobile Used the same binary?
02:25 [mbm] yeah, dunno what went wrong before
02:25 [mbm] the zipinfo size was correct but the file size was actually much larger
02:36 [mbm] ah, figured out what happened
02:39 CEnnis91|Mobile what was is
02:40 [mbm] zip is stupid when the zip file already exists
02:41 CEnnis91|Mobile Did it include the original zip?
02:41 [mbm] no, but for some reason it didn't remove old zip data even though it apparently updated the index not to list those file
02:43 CEnnis91 [mbm]: so i ended up giving ubuntu 15gb for / and the rest to /home, think ill have any issues down the road?
02:44 [mbm] heh, I've been running a version of ubuntu stuffed into 4g :P
02:44 CEnnis91 im sure i run quite a few more gui applications compared to you however =P
02:44 CEnnis91 after initial install and download of the basics, i have 28% of that used
02:45 [mbm] also I usually avoid splitting the filesystem into several partitions because inevitably one will fill faster than the others
02:45 CEnnis91 yea i just did / and /home
02:45 [mbm] although I suppose that's somewhat moot when you can just reboot into a live cd and move partitions around
02:45 CEnnis91 how does that work without data loss?
02:46 CEnnis91 i never understood that, and didnt wanna risk data trying
02:46 [mbm] benefit of splitting home means that you can keep it while reformatting everything else if you reinstall
02:46 [mbm] you mean how does resizing a partition work?
02:47 CEnnis91 yea thats why i decided to split my partitions up, i dont want to have to deal with data transfer again on this laptop
02:47 [mbm] making a partition bigger is easy, just update a few bytes in the filesystem header that indicate the total size
02:47 CEnnis91 i suppose so, but can you really just resize one and append to another?
02:48 CEnnis91 yea, well i usually fill the entire hdd
02:48 CEnnis91 so i would be making smaller, not larger
02:48 [mbm] making a partition smaller is pretty much defragmenting so that the free space is all in one chunk at the end
02:48 [mbm] and then adjusting the headers
02:50 CEnnis91 does that need to be done manually or is it done when you say "make smaller"
02:51 [mbm] it's all automatic; you just drag around a linear chart representing the partitions in gparted
02:51 [mbm] which btw is a blatent knockoff of an older dos utility called partition magic
02:56 CEnnis91 heh
02:56 CEnnis91 yea i wouldnt be old enough to recognize knockoffs
02:59 CEnnis91 well shit, the panel applet i used in 11.10 doesnt work in 12.04
03:00 CEnnis91 oh nvm, someone is just a retard
03:14 CEnnis91 [mbm]: considering you mentioned using ubuntu, did you bother to get into the software center crap that lets you sync across devices?
03:16 [mbm] I just use network shares
04:24 CEnnis91 oh cool android updated the build environment for 12.04
04:24 CEnnis91 copy,paste,done
04:24 CEnnis91 although im pretty sure openjdk doesnt work well
04:24 CEnnis91 especially when eclipse gets involved
04:25 CEnnis91 [mbm]: im gonna guess you know what ccache is right?
04:26 CEnnis91 oh nvm, i answered my own question
05:11 christefano joined #motoactv
07:43 [mbm] CEnnis91: yeah, ccache is fun
07:43 [mbm] distcc is also fun
07:50 CEnnis91 yea i realized my computer isnt even close to being able to handle it
07:50 [mbm] ccache? just takes a bit of disk space
07:51 CEnnis91 yea, which i dont have really
07:51 CEnnis91 i think i saw 50gb?
07:51 [mbm] well, it'll use as much disk space as you throw at it
07:52 CEnnis91 damnit
07:52 CEnnis91 ./dev/sda6        15G  5.5G  8.2G  40% /
07:52 CEnnis91 udev            2.0G   12K  2.0G   1% /dev
07:52 CEnnis91 tmpfs           784M  940K  783M   1% /run
07:52 CEnnis91 none            5.0M     0  5.0M   0% /run/lock
07:52 CEnnis91 none            2.0G  1.6M  2.0G   1% /run/shm
07:52 CEnnis91 ./dev/sda7       111G   16G   90G  15% /home
07:52 [mbm] all it's doing is caching the compiler output so it doesn't have to run the compiler commands again
07:52 CEnnis91 where would you like me to put it, lol, im still downloading cm
07:53 CEnnis91 yea i get what it does
07:53 CEnnis91 is doing a small amount of cache even worth it?
07:53 CEnnis91 say 25gb?
07:54 [mbm] starts to become worthless if the ccache size is smaller than the space required to compile everything in your project
07:54 CEnnis91 thought so
07:54 CEnnis91 im guessing thats why they suggested 50
07:55 CEnnis91 anything else that can speed up a compile?
07:55 CEnnis91 (ramdisk wouldnt be plausible for me either)
07:56 [mbm] well, start by using make's -j option to specify the number of jobs to run in parallel
07:56 CEnnis91 yea, i can handle about -j4
07:56 CEnnis91 before my computer becomes completely unusable
07:56 [mbm] minimum for -j is the number of cores you have
07:56 CEnnis91 2
07:57 [mbm] then yeah, you're maxed out at -j4
07:57 [mbm] (each core having one job running and another waiting)
07:57 CEnnis91 ah, makes sense
07:57 CEnnis91 is that improved with i5-i7?
07:57 CEnnis91 or it would still be 2x # of cores?
07:58 CEnnis91 (Intel® Core™2 Duo CPU T6600 @ 2.20GHz × 2 -- fwiw)
07:58 [mbm] still pretty much 2x cores
07:59 [mbm] although some people still use cores+1
08:00 CEnnis91 what if you went beyond 2x cores? just no performance? or actually negative performance?
08:00 CEnnis91 no performance boost*
08:01 [mbm] cores+1 with higher core counts can result in minor gaps where a core is doing nothing just because the job queue is empty
08:01 [mbm] well, you can actually leave the number off -j
08:01 [mbm] at which point it becomes a fork bomb
08:02 CEnnis91 fork bomb?
08:02 [mbm] and completely cripples the computer
08:02 CEnnis91 good to know
08:02 CEnnis91 -j = run as fast as you want?
08:02 [mbm] a fork bomb is a program which runs several copies of itself
08:02 CEnnis91 oh
08:03 [mbm] which basically means that the computer keeps running more and more until it crashes
08:03 [mbm] ie, eventually it runs out of ram and something dies
08:03 CEnnis91 and the reason that this hasnt been fixed?
08:04 [mbm] hmm? there's nothing to fix
08:04 CEnnis91 so make -j is by design meant to do that?
08:04 [mbm] well, -j means that there's no limit on the number of jobs that can be run in parallel
08:04 CEnnis91 i would guess if it was a small make with sufficient ram, you would get the best performance?
08:05 CEnnis91 okay
08:05 CEnnis91 so i was right before
08:05 [mbm] that's not to say that the environment doesn't influence it
08:06 [mbm] if the project consists of files a b and c, each depending on the previous then there's no way it can ever run more than one at a time
08:07 CEnnis91 ah
08:07 [mbm] but when you get into something like android, there's always a million things things that can be run
08:07 CEnnis91 and im guessing you're leading to the point that because android is a cluster, its giong to fail if compiled this way?
08:08 [mbm] and the problem with -j there is a combination that there's always going to be a million things in the queue, along with the computation of the dependencies for those million things
08:10 CEnnis91 yea, makes sense
08:10 [mbm] something else that should be considered is the physical effects of keeping the cpu completely loaded for a prolonged period of time
08:11 [mbm] it generates a lot of heat, and not all systems are designed to handle the cooling
08:11 CEnnis91 well yea, im pretty much fucked at that point
08:11 CEnnis91 im surprised this thing is still running
08:12 [mbm] cpu will save itself by going into thermal throttle if it gets too hot
08:12 [mbm] basically reducing the effective MHz of the cpu and inserting a ton of do-nothing commands to give the cpu a chance to cool down
08:13 [mbm] but you can still do a reasonable bit of damage to the rest of the system
08:13 CEnnis91 i meant more for consistently running it at full throttle
08:14 CEnnis91 the only thing seeming to go is the screen, which im pretty sure is unrelated to anything ive done in that respect
08:14 CEnnis91 and by seeming, i mean, it fades to white consistently
08:15 [mbm] screen goes a bit cloudy and then fades to white a bit unevenly?
08:16 CEnnis91 i think so
08:16 CEnnis91 basically like the power is being cut
08:16 CEnnis91 i get vertical colored lines
08:16 CEnnis91 almost positive its the wiring in the hinges
08:18 [mbm] well, with an lcd the backlight is on constantly
08:19 [mbm] the lcd crystals that make up a pixel twist the polarization and you end up with black
08:20 CEnnis91 well that would explain the white
08:20 [mbm] basically you have to charge the pixel to make it black
08:21 [mbm] and often you have to keep refreshing
08:21 CEnnis91 seems energy deficient
08:22 [mbm] well, either the graphics driver manually refreshes or the lcd controller supports an auto refresh
08:23 [mbm] when it stops getting refreshed the pixels lose their charge
08:23 [mbm] and the fade to white pattern is each pixel individually discharging
08:23 CEnnis91 so either the driver isnt sending the signal, or the signal isnt making it
08:24 [mbm] usually different parts of the screen discharge at a slightly different rate
08:24 CEnnis91 usually goes top down
08:25 [mbm] and yeah, you're completely right, either the refresh signal isn't generated or it's not making it all the way to the screen
08:25 CEnnis91 well how do the colors form
08:26 CEnnis91 colored vertical lines
08:26 [mbm] which can also happen if you have a bad cable
08:26 CEnnis91 well because it happens when the screen angle is adjusted, it makes me think its the cable/wires
08:28 [mbm] are we talking a few stray lines or is there a pattern to it?
08:28 CEnnis91 stray
08:29 [mbm] :/
08:29 CEnnis91 whats that mean
08:31 [mbm] well, you have a matrix of pixels that need to be charged
08:31 [mbm] and rather than running a separate line to each pixel
08:32 [mbm] they get arranged into a row/column
08:33 [mbm] but when the wiring fails, it means that suddenly you lose a whole row or column
08:33 [mbm] which then discharges and leaves you with a line
08:34 CEnnis91 but why colored?
08:34 CEnnis91 it discharges midway through the signal?
08:35 [mbm] well, to say that the lcd is actually controlled at a pixel level is a bit misleading
08:35 [mbm] it's actually r g b of each pixel
08:35 CEnnis91 yea i figured that
08:35 CEnnis91 considering the majority of the lines are either r g or b
08:36 CEnnis91 occasionally the blend of 2
08:37 [mbm] for layout reasons, the control lines follow a strict horizontal or vertical
08:37 [mbm] and basically control whatever they pass
08:38 [mbm] the backlight is white, and somewhere between you and the backlight are colored filters
08:38 [mbm] to tint the color of the backlight
08:38 CEnnis91 random: omg i just noticed cm7 has 313 packages...guess its time to make a custom manifest
08:39 [mbm] different vendors like different arrangements for where to place the rgb
08:39 [mbm] some will line up all the reds, all the greens, all the blues
08:39 [mbm] and others favor a more checkerboard pattern
08:40 CEnnis91 why not just stick to one template?
08:41 [mbm] well, the eye is particularly bad at distinguishing blue
08:41 [mbm] so does each pixel need a blue?
08:41 [mbm] or can I cheat and just leave blue off half the pixels?
08:42 [mbm] it's that sort of thinking that results in things like pentile
08:42 CEnnis91 so money, shouldve known
08:42 CEnnis91 well and battery i guess
08:43 [mbm] but yeah, basically comes down to the contorl line controlling exactly what it passes
08:43 [mbm] and depending on the rgb layout it can be either a single color or alternating set of colors
08:44 CEnnis91 alright, so then i have another question, why are certain columns the only ones that show up
08:45 [mbm] eg, if some vendor decided that their screen was only a checkerboard of red and green then a line through it would be an alternating pattern of red/green
08:45 CEnnis91 like one 2 inches from the left is always green
08:47 [mbm] so, from the lcd controller to the panel itself you have one wire per column
08:47 CEnnis91 oh really?
08:47 [mbm] and the one 2 inches from the left is the one that happened to fail
08:47 CEnnis91 yea that answered my question right away
08:47 CEnnis91 this isnt very nice to fix either right?
08:48 [mbm] it's part of a ribbon cable
08:49 [mbm] if you've ever taking apart any electronics you've probably seen them, a thin orangish plastic
08:49 Snoothy joined #motoactv
08:50 CEnnis91 yes i know what that is...in a basic sense
08:50 CEnnis91 i would just need to replace that?
08:50 [mbm] http://techon.nikkeibp.co.jp/engli​sh/NEWS_EN/20070919/139324/CCC.jpg
08:51 [mbm] there's an example
08:51 CEnnis91 yep
08:51 [mbm] the one on the left you can see has the lcd controller integrated into the ribbon
08:51 [mbm] so there are few wires going in
08:52 [mbm] then after the controller theres a ribbon that's suspiciously as wide as the lcd
08:53 [mbm] and for obvious reasons, that ribbon is bonded directly to the glass
08:54 [mbm] with the glass itself having a conductive pattern that carries the signal to the subpixels
08:54 CEnnis91 so basically the one thing i need to replace is a pita to replace?
08:54 CEnnis91 and likely expensive
08:54 [mbm] and the failure point is almost always where the ribbon meets the glass
08:55 [mbm] and while you can add a bit of pressure to that point to temporarily fix it, you're putting strain on the rest
08:56 CEnnis91 =/
08:56 [mbm] and yeah, you're pretty much talking replacement of the panel assembly
08:56 CEnnis91 putting pressure on the hinges can usually fix it (of course temp)
08:56 CEnnis91 so thats what i do sometimes
08:59 [mbm] some small percentage of lcd panels just fail by themselves as whatever glues the ribbon to the lcd comes undone
09:00 [mbm] but usually the failure is physical stress
09:00 [mbm] a laptop that's been dropped or had something heavy sitting on it
09:00 CEnnis91 yea i would attribute it to heavy usage
09:00 CEnnis91 probably the way ive carried it
09:00 CEnnis91 had this for 3 years now
09:00 CEnnis91 and the build quality was eh
09:01 [mbm] yeah, pretty common in the used laptop market
09:01 CEnnis91 lol i got it new -_-
09:01 [mbm] well, I mean that by the time a laptop becomes a used laptop the screen has failed
09:02 CEnnis91 ah, yea
09:03 [mbm] good laptops have an all metal case
09:03 [mbm] and the cheap ones are flexable plastic
09:03 CEnnis91 oh yea this was cheap and plastic
09:03 CEnnis91 inspiron 1545
09:04 [mbm] lol, dell
09:04 CEnnis91 yea i didnt have much choice
09:04 CEnnis91 my gf gave me an old aspire one
09:05 CEnnis91 i picked it up and the build quality was somuch better
09:05 JohnDoe_ joined #motoactv
09:05 [mbm] you basically have the choice of buying something cheap and wearing it out quickly or buying something more expensive that will last 5+ years
09:06 CEnnis91 my mom isnt in that mindset at all
09:07 [mbm] well, most people don't see it that way
09:07 CEnnis91 she still uses an old dell that shipped with xp (desktop) -- i wanna say 512mb  ram?
09:07 [mbm] and to be fair, how long does it need to last before it's obsolete?
09:07 CEnnis91 in whos eyes?
09:07 CEnnis91 to some its obsolete before its released
09:08 [mbm] well, I mean some of the computer equipment they built in the 80s is strong as a tank and works just as good as when it was made
09:08 [mbm] good solid build quality, but who wants something from 20+ years ago?
09:09 CEnnis91 oh yea exactly
09:09 [mbm] it's actually possible to go overboard on build quaility to the point it becomes a waste of money
09:10 CEnnis91 yea for a general consumer sure
09:10 CEnnis91 but there are some who will benefit greatly fromit
09:10 [mbm] right
09:11 [mbm] there's a whole market designed for people who just want something cheap to last them until the next technological innovation
09:11 [mbm] which also appeals to those who are simply influence by the fact it's cheap
09:12 CEnnis91 really? theres still companies that do the excellent build quality on products?
09:12 CEnnis91 like
09:12 CEnnis91 beyond excellent
09:13 Snoothy lenovo/ibm ?
09:13 CEnnis91 it only seems like shit thats made like a tank is made20 years ago, like you mentioned
09:13 [mbm] well, I'm no apple fanboy but damn they build some quality hardware
09:13 CEnnis91 yea
09:13 [mbm] ibm is also good, although technically that's lenovo
09:14 [mbm] used to be subcontracted out and sold under the ibm brand
09:14 [mbm] and now lenovo is trying to make a name for themselves by selling directly
09:15 CEnnis91 i could be wrong
09:15 [mbm] you'll actually find that's true of a lot of electronics; the brand name associated with it is simply a distributor
09:15 CEnnis91 but dont you own a lenovo laptop
09:15 [mbm] me? several
09:15 CEnnis91 i feel like you;ve mentioned being on one
09:15 CEnnis91 ah
09:16 [mbm] just bought another earlier this week to replace an aging eeepc701
09:17 * [mbm] loves the thinkpad line of laptops
09:17 CEnnis91 but does lenovo have performance computers, or just reliable?
09:17 Snoothy it's mostly reliable business computers IMO
09:17 [mbm] the one I just bought is a cheap ($500-ish) thinkpad x310e
09:19 CEnnis91 about the specs of my computer now
09:19 CEnnis91 except updated proc
09:20 CEnnis91 looking these arent something i would want to build cm on
09:20 CEnnis91 thats for sure
09:21 [mbm] Snoothy: well, there are other companies that are just as reliable; what I like about the thinkpads is just that you can pick pretty much any of them and it'll be good, vs one or two good ones and a minefield of crap
09:22 CEnnis91 oh the higher models do have i5/7
09:23 [mbm] I bought this one mostly just on the basis that it was cheap and I had some newegg credit laying around
09:23 CEnnis91 lol
09:24 [mbm] apparently it's part of their education series of student laptops
09:24 CEnnis91 finally cm7 finished syncing
09:24 [mbm] so it's cheap plastic but it has some heft to it
09:24 CEnnis91 $500 isnt bad i suppose
09:25 CEnnis91 if this ever goes
09:25 CEnnis91 whats the battery life
09:25 [mbm] can't say I much care for the design, but the build quality is there
09:25 CEnnis91 yea design is meh all around from what i saw
09:26 [mbm] the back of the screen is rounded like a dome
09:26 [mbm] which you can't tell in the pictures
09:26 [mbm] makes for a stronger screen, but design wise it looks like crap
09:27 [mbm] I'm also not a particular fan of the chicklet keyboards although that seems to be all you can buy in new laptops
09:28 [mbm] ie, every key is an individual pushbutton with a gap between the keys
09:28 CEnnis91 i dont mind them
09:29 CEnnis91 meh alright i need to get at least a few hours of sleep
09:29 [mbm] they're not bad but the typing feel is different
09:30 [mbm] haven't used the chicklet style long enough to know if taking the keyboard apart will be a problem
09:31 [mbm] my eeepc701 I physically wore out the keyboard to the point that a number of keys would nolonger work
09:31 [mbm] each key has little rubber domes under it
09:32 [mbm] and I'd started taking the keyboard apart and shuffling them around
09:33 [mbm] wore out the w key, took the keyboard apart, swapped parts around so now the printscreen button doesn't work
09:33 [mbm] but who the hell needs that key
09:33 JohnDoe_ hat, hen, here?
09:35 [mbm] several rounds of similar replacements
09:35 [mbm] basically I kept putting the worn parts on the extended keys that I didn't use
09:35 [mbm] printscreen, windows keys, right alt/ctrl ..
09:36 [mbm] press any of those and there's no click, key just completely caves in, doesn't even register
09:39 [mbm] let's see, the eeepc701 was from 2007
09:40 [mbm] not bad .. still works, although the keyboard and mouse buttons are nearly worn out
09:41 [mbm] 600Mhz celeron and a 640x400 screen are somewhat crippling
09:48 kr1t1c4l eee were good in their day; iirc they helped the netbook trend
09:50 [mbm] yep, netbooks survived, low cost no cdrom laptops .. but the under 11" form factor seems pretty much dead
09:51 [mbm] shame too; reason I liked the eeepc was that it was the perfect size for an ssh/irc terminal
10:50 christefano joined #motoactv
12:15 kr1t1c4l- joined #motoactv
19:02 christefano joined #motoactv
20:53 Quietlyawesome94 joined #motoactv
21:39 christefano joined #motoactv
21:44 christefano joined #motoactv
22:01 christefano joined #motoactv
22:42 [mbm] CEnnis91: what's that xda post of yours about not being able to return 16g to stock?
22:42 CEnnis91 i never actually released the fxz yet
22:42 [mbm] oh
22:43 CEnnis91 he just didnt seem phased by the fact that motocast thought he had an 8gb
22:43 CEnnis91 i had to say something
22:44 [mbm] btw, filesystems usually compress better if you start with the entire thing as 0x00, format and then copy files once
22:45 [mbm] that way all the unused space is still 0x00 instead of old data
22:46 CEnnis91 ummm
22:46 CEnnis91 oh
22:47 CEnnis91 meaning write everyting to 0x00, format partition, push files in, pull .img, compress?
22:47 [mbm] yes, although you can do it with a file on your pc instead of reformatting the device
22:48 CEnnis91 now im lost, you're going to have to explain how that is done
22:48 CEnnis91 i thought .img's couldnt be made
22:49 [mbm] sure, why not? they're not signed or anything
22:49 [mbm] dd if=/dev/zero of=disk.img bs=1M count=200; mkfs.ext2 disk.img; mount -o loop disk.img ...
22:50 ds2 .img's got to come from somewhere!
22:50 ds2 we don't have spontaneous .img creation :D
22:50 CEnnis91 [mbm]: then mount -o loop old.img; cp old.img disk.img?
22:51 CEnnis91 or should it be moved differently
22:51 [mbm] close enough
22:51 CEnnis91 well yea the cp is wrong, but the idea
22:52 [mbm] from a technical point of view 0xff might be a better use of unused space than 0x00
22:52 CEnnis91 because thats the real state of a blank partition?
22:53 CEnnis91 [mbm]: oh, maybe its that i was told you cant edit an .img?
22:53 CEnnis91 or is that too possible
22:54 [mbm] well, a flash chip will have something like a 64k (example) erase size, when erased everything goes to 0xff
22:54 [mbm] erll, everything in that 64k
22:54 [mbm] writing a byte inside that changes some of the 1's to 0's
22:54 CEnnis91 yep
22:55 CEnnis91 ive heard you explain this one before =)
22:55 ds2 PhD paper idea -
22:55 [mbm] but as soon as you have a 0x00 and you want to change it to ssomething else you're stuck erasing the whole 64k
22:55 ds2 typical flash life and speed study of the difference of direct flash vs flash with an inverter in front of it
22:57 [mbm] so leaving the unused space as 0xff may save you a few erase cycles
22:57 [mbm] at least, theoretically; the driver might be stupid and do them anyway ;)
22:58 CEnnis91 haha
22:58 CEnnis91 [mbm]: now how much space would i really save by doing this?
22:58 [mbm] although given that flash chips usually have 100,000-1,000,000 cycle life expectency you'd really need to be overwriting a lot before you broke anything
22:58 CEnnis91 is it worth it, or were you just stating, for the hell of it
22:59 ds2 that's for SLC chips
22:59 [mbm] well, like I said before my repack got down to 162M
22:59 ds2 it can be a lot worse for MLC
22:59 CEnnis91 [mbm]: oh, hmmm
23:00 [mbm] ds2: yep, and I'm not counting spare
23:00 CEnnis91 well either way i still need to finish setting up ubuntu again -_-
23:00 CEnnis91 so tonight it's not gonna happen
23:02 [mbm] also, your comment in regards to modifying img files .. what do you think the mount command after the mkfs was doing? :)
23:06 CEnnis91 yea, lol idk what i was told then
23:34 Quietlyawesome94 joined #motoactv

| Channels | #motoactv index | Today | | Search | Google Search | Plain-Text | summary