**** BEGIN LOGGING AT Thu Dec 19 02:59:59 2013
Dec 19 08:07:37 <kerio> hbib: the entire point of the whole exercise is to have something that runs faster
Dec 19 08:07:43 <kerio> how could it run slower?
Dec 19 08:15:23 <hbib> my understanding of thumb was that it frees RAM, but needs a little longer to execute a command because of the correction mechanisms for the ARM Erratas
Dec 19 08:24:02 <DocScrutinizer05> hbib: that's basically correct
Dec 19 08:24:58 <DocScrutinizer05> but the saved RAM overcompensates for the slower command execution, particularly when device runs into swapping (which it basically does all the time)
Dec 19 08:35:57 <hbib1> DocScrutinizer05: So thumb runs all in all faster? i didn't tried it yet because i was always "afarid" of the cpu overhead ...
Dec 19 08:37:15 <DocScrutinizer05> the overhead is not really noticeable, but I must admit I also haven't done proper evaluation of the over-all speed increase from thumb, I just tested browser and it felt quite a bit faster than usual
Dec 19 08:37:18 <DocScrutinizer05> see
Dec 19 08:37:22 <DocScrutinizer05> ~thumb
Dec 19 08:37:22 <infobot> methinks cssu-thumb is <DocScrutinizer05> [thumb2 microb] indeed seems to render like mad, subjectively, or http://talk.maemo.org/showpost.php?p=1220597, or http://wiki.maemo.org/Community_SSU/Thumb
Dec 19 08:38:25 <DocScrutinizer05> my major concern is with evaluation/QA of thumb-recompiled apps
Dec 19 08:39:01 <DocScrutinizer05> though we haven't seen a thumb-induced bug yet, afaik
Dec 19 08:41:01 <DocScrutinizer05> the other main problem is that thumb binaries will make a non-thumb-enabled system segfault to death (I.E. a system that doesn't have the patched kernel, whether that's thuumb-kernel or the recent KP)
Dec 19 08:44:02 <DocScrutinizer05> hmm, it just occurred to me if cssu-thumb could use unionfs to mount /opt/usr/bin-thumb/ over /usr/bin and thus replace the arm binaries by their thumbified equivalents. Then restart what needs a restart - e.g. browserd
Dec 19 08:45:30 <DocScrutinizer05> this way even a system with a non thumb enabled kernel could start up and even work normally when the process that does the unionfs mount checks if the kernel actually supports thumb-workaround
Dec 19 08:46:00 <DocScrutinizer05> where's freemangordon when I have something he should notice and consider? :-P
Dec 19 08:47:51 <hbib1> if i understand you correct, that would be a nice idea, because it would be possible to switch between thumb and non-thumb without reflashing
Dec 19 08:48:05 <DocScrutinizer05> basically yes
Dec 19 08:48:13 <hbib1> and so it would be possible to just test thumb without consequences ...
Dec 19 08:48:22 <DocScrutinizer05> :nod:
Dec 19 08:49:07 <DocScrutinizer05> this particular thumb concept could get installed on every system and wouldn't cause any major trouble, since it kicks in only when it detects it can work
Dec 19 08:50:10 <DocScrutinizer05> thus you can install (and of course deinstall) it on arbitrary fremantle systems
Dec 19 08:52:48 <hbib1> that would be nice!
Dec 19 08:53:01 <DocScrutinizer05> you'd mount the unionfs in three layers, like /usr/bin over /opt/usr/bin-thumb over (original) /usr/bin, to allow normal updates of apps in /usr/bin. For thumb binary updates you'd update /opt/user/bin-thumb/* rather than /usr/bin
Dec 19 08:54:36 <DocScrutinizer05> of course the thumb-enabler-daemon could also just bindmount each single thumb binary over its arm counterpart in (/usr)/bin, thus no unionfs needed
Dec 19 08:55:16 <DocScrutinizer05> but could tend to get messy in output of mount(8)
Dec 19 08:56:34 <DocScrutinizer05> .
Dec 19 08:58:03 <DocScrutinizer05> and the simplest approach is to add /opt/bin-thumb:/opt/sbin-thumb to $PATH
Dec 19 08:58:50 <DocScrutinizer05> this however would require all other stuff like .desktop files etc to correctly use $PATH concept
Dec 19 09:02:56 <hbib1> i like the unionfs-idea, because i think it is the easiest and most transparent way ...
Dec 19 09:11:13 <DocScrutinizer05> sure
Dec 19 09:11:28 <DocScrutinizer05> just we lack unionfs drivers on maemo afaik
Dec 19 09:11:45 <DocScrutinizer05> of course thumb-package could pull them in
Dec 19 09:12:35 <DocScrutinizer05> I also sorted the alternatives by decreasing appeal to myself
Dec 19 09:12:39 <DocScrutinizer05> ;-)
Dec 19 09:41:51 <kerio> DocScrutinizer05: there's not enough space though
Dec 19 09:41:59 <DocScrutinizer05> huh?
Dec 19 09:42:14 <DocScrutinizer05> then make /opt larger
Dec 19 09:42:31 <kerio> emmc is slow
Dec 19 09:42:36 <DocScrutinizer05> tzz
Dec 19 09:42:42 <DocScrutinizer05> says who?
Dec 19 09:43:28 <DocScrutinizer05> btw to start a few instances of e.g. browserd it's absolutely irrelevant when it takes twice as long
Dec 19 09:47:01 <DocScrutinizer05> IroN900:~# dd if=/dev/mtd5 of=/dev/null bs=10M count=10
Dec 19 09:47:02 <DocScrutinizer05> 10+0 Datensätze ein
Dec 19 09:47:04 <DocScrutinizer05> 10+0 Datensätze aus
Dec 19 09:47:05 <DocScrutinizer05> 104857600 Bytes (105 MB) kopiert, 5,74506 s, 18,3 MB/s
Dec 19 09:47:08 <DocScrutinizer05> IroN900:~# dd if=/dev/mmcblk0p2 of=/dev/null bs=10M count=10
Dec 19 09:47:10 <DocScrutinizer05> 10+0 Datensätze ein
Dec 19 09:47:11 <DocScrutinizer05> 10+0 Datensätze aus
Dec 19 09:47:13 <DocScrutinizer05> 104857600 Bytes (105 MB) kopiert, 6,12222 s, 17,1 MB/s
Dec 19 09:47:43 <DocScrutinizer05> yeah, emmc is slow
Dec 19 09:47:48 <DocScrutinizer05> just like NAND is
Dec 19 09:49:45 <DocScrutinizer05> kerio: any other questions, Watson?
Dec 19 09:50:08 <kerio> why do you have to be so mean tho :c
Dec 19 09:50:16 <DocScrutinizer05> LOL
Dec 19 09:50:21 <jaska> born that way?:)
Dec 19 09:50:28 <DocScrutinizer05> sure
Dec 19 09:50:41 <kerio> i still disapprove of having to keep two sets of binaries
Dec 19 09:50:50 <kerio> especially if you consider that you have to keep them both updated
Dec 19 09:51:07 <DocScrutinizer05> errr
Dec 19 09:51:47 <DocScrutinizer05> with 32GB of storage, I prefer 2 sets of binaroes over a system that worst case doesn't boot anymore
Dec 19 09:52:04 <ShadowJK> It's more that writes on the emmc blocks reads, not that reads in themselves are slow
Dec 19 09:52:18 <DocScrutinizer05> :nod:
Dec 19 09:52:31 <DocScrutinizer05> which is irrelevant for the idea sketched above
Dec 19 09:52:40 <ShadowJK> how so?
Dec 19 09:52:51 <DocScrutinizer05> how not?
Dec 19 09:53:20 <DocScrutinizer05> loading and starting binaries rarely involves any writing to storage
Dec 19 09:53:52 <kerio> except when you're swapping on the emmc itself
Dec 19 09:54:49 <DocScrutinizer05> you're not going to swap anything more when (re!)starting a binary from eMMC than you have to do when starting same binary (though ARM and thus larger) from NAND
Dec 19 09:56:02 <DocScrutinizer05> I rather expect to see RAM getting freed up than to have anything swapped out, when killing ARM browserd and starting thumb browserd from eMMC
Dec 19 09:56:06 <ShadowJK> binaries are never loaded in one go, it's all read in on-demand, so if you access a functionality you haven"t touched before in that session (or one that was ejected from ram as it hadnt been used for long), it needs to be read back in. There's all sorts of write activity going on all the time, type a URL in the browser or tap a link can cause up to several seconds of emmc stalls just from the browser history and typed URLs getting updated :)
Dec 19 09:56:30 <jon_y> depends
Dec 19 09:56:37 <jon_y> if it fits in a page
Dec 19 09:56:46 <jon_y> the whole blob is loaded
Dec 19 09:56:53 <DocScrutinizer05> ShadowJK: when that were 100% true, then preloading of e.g. browserd makes no sense at all
Dec 19 09:57:28 <zamn900> too boring and too long for me
Dec 19 09:57:31 <ShadowJK> browser preloading also saves the cpu cycles on load
Dec 19 09:57:32 <zamn900> oh gosh
Dec 19 09:58:10 <jon_y> so is all preloaded data executable code?
Dec 19 09:58:14 <ShadowJK> and coincidentally pulls in lots of library and executable, that later gets thrown out again if there's memory pressure :)
Dec 19 09:58:33 <jon_y> I can imagine they contain assets that would otherwise be manually loaded later
Dec 19 09:58:51 <ShadowJK> jon_y; alot of the boost comes from lots of initialization being done in advance
Dec 19 09:59:08 <jon_y> you mean in the background?
Dec 19 09:59:12 <ShadowJK> yes
Dec 19 09:59:32 <jon_y> Windows and IOS is doing this a lot
Dec 19 09:59:45 <DocScrutinizer05> ShadowJK: that's all maybe correct (or maybe not applicable in a number of cases) but highly irrelevant when kerio complains about eMMC being too slow for holding thumb binaries to replace the ARM binaries used during boot from rootfs/NAND
Dec 19 10:00:20 <ShadowJK> jon_y; that's sort of a different thing
Dec 19 10:01:18 <ShadowJK> DocScrutinizer05; i'd do it another way, hold a thumb kernel on emmc so if user despite warnings install non-thumb kernel, there'd be backup one..
Dec 19 10:01:54 <DocScrutinizer05> which is pretty useless unless you got a uboot kernel
Dec 19 10:01:58 <jonwil> hi
Dec 19 10:02:20 <DocScrutinizer05> and when you have requirements on the kernel then we are where we started
Dec 19 10:02:36 <DocScrutinizer05> require,emt: kenel supports thumb
Dec 19 10:02:47 <DocScrutinizer05> requirememt: kernel supports thumb
Dec 19 10:02:57 <DocScrutinizer05> otherwise system won't boot
Dec 19 10:03:14 <DocScrutinizer05> I don't see how a second kernel on eMMC can cure this
Dec 19 10:03:19 <ShadowJK> I don't like the idea of often used stuff residing on emmc. Too big of a performance penalty :/
Dec 19 10:03:43 <DocScrutinizer05> how is that a performance penalty?
Dec 19 10:04:03 <jonwil> Don't expect any more N900/Neo900 work from me anytime soon, my N900 dev box went to the e-waste facility (it was too old to be worth keeping)
Dec 19 10:04:15 <DocScrutinizer05> actually 50..95% of our binaries live in /opt on eMMC
Dec 19 10:04:39 <ShadowJK> Can see this with fennec, with fennec, ~/.mozilla/fennec and swap on emmc, it's much much slower than stock browser, but after splitting it up to rootfs, emmc and uSD, it runs about 5 times faster :D
Dec 19 10:04:58 <DocScrutinizer05> jonwil: would sponsoring a new box help ?
Dec 19 10:06:13 <jonwil> Not sure that I have much in the way of time to devote either, I will be looking for a new apartment soon (staying with family for now) and I also need to focus a lot more on finding someone to, ya know, pay me to do stuff :)
Dec 19 10:06:31 <DocScrutinizer05> ShadowJK: this doesn't sound like a valid argument against my suggestion to use unionfs to switch from plain maemo5 to thumbified maemo5, *after* system boot
Dec 19 10:07:24 <DocScrutinizer05> jonwil: a pity! we#ll miss your help
Dec 19 10:07:32 <ShadowJK> thumb and normal system on emmc, choose which one to copy to rootfs on boot or what?
Dec 19 10:07:46 <DocScrutinizer05> see backscroll please
Dec 19 10:08:18 <DocScrutinizer05> starting at [2013-12-19 09:44:02] <DocScrutinizer05> hmm, it just occurred to me 
Dec 19 10:08:25 <jonwil> Not sure exactly what n900 work I could do anyway
Dec 19 10:09:21 <jonwil> If it wasn't for the big $ I am spending on this whole moving house thing, I would be upgrading my main box (Windows 7 PC) to a Core i5 or i7 and turning this Core 2 Duo into a Linux box. But I am broke
Dec 19 10:09:44 <ShadowJK> Yeah I don't like it
Dec 19 10:10:04 <ShadowJK> I want stuff removed from emmc, not added to it :D
Dec 19 10:10:21 <DocScrutinizer05> remove emmc then!
Dec 19 10:10:26 <DocScrutinizer05> :-P
Dec 19 10:10:36 <ShadowJK> That would be the ultimate goal.
Dec 19 10:10:57 <ShadowJK> It's a tape drive, only good for storing huge files :P
Dec 19 10:11:14 <DocScrutinizer05> *shrug*
Dec 19 10:11:38 <DocScrutinizer05> I don't see anything better around the corner. and for sure not readily available, for embedded
Dec 19 10:12:15 <ShadowJK> Though, with a new kernel, and f2fs instead of ext3, it'd be usable
Dec 19 10:13:09 <ShadowJK> Soldering on a sandisk extreme microsdhc, if it was possible, would be on the order of 3-4 magnitudes improvement actually
Dec 19 10:13:34 <jonwil> so mostly its my lack of dev box and my lack of motivation to do n900 work that is my problem right now
Dec 19 10:14:56 <DocScrutinizer05> jonwil: while we possibly could help with the former, the latter is a killer argument
Dec 19 10:15:51 <DocScrutinizer05> when Neo900 and fremantle porting is no motivation to you, I'm afraid we can't find anything to motivate you
Dec 19 10:17:15 <jonwil> I suspect if I had a specific narrow target to focus on, my motivation would come back.My problem right now is that I keep jumping from binary to binary looking for things to reverse engineer, trying one, giving up because its too hard, trying another one, etc
Dec 19 10:18:22 <DocScrutinizer05> maybe your goals are to hard to reach, so the joy of achieving them is what lacks
Dec 19 10:19:17 <DocScrutinizer05> too hard*
Dec 19 10:20:37 <zamn900> I said is rto
Dec 19 10:20:40 <zamn900> too
Dec 19 10:24:21 <jonwil> I do wonder if anyone has done anything notable with the MCE code yet
Dec 19 10:25:08 <jonwil> doesn't look like it from what I can see :)
Dec 19 10:26:10 <DocScrutinizer05> anyway I have a todo list the length of a roll of toilet paper, for today. appointment with my med. doc, searching for info about founding a company (actually a "limited"), writing some letters to municipal office on related stuff, tax office to get a new company tax number, writing some review of Neo900 lower board mech evaluation, open letter ("to FSF") about Neo900 privacy concept and infeasibility to comply with some particular rules 
Dec 19 10:26:11 <DocScrutinizer05> regarding FSF and their take on a free system, writing a huge spec about what to implement in which way on Neo900 hardware and related on FPTF
Dec 19 10:26:54 <jonwil> I will certainly contribute inputs on the FPTF threat on thoughts about how to do things
Dec 19 10:27:20 <DocScrutinizer05> some people are using it, one bug been spotted iirc, I published a patch (binary edit for now) for indicator LED which pali/fmg want to implement to CSSU
Dec 19 10:27:42 <DocScrutinizer05> (MCE that is)
Dec 19 10:28:10 <DocScrutinizer05> fixing the engine3 idiocy is on my ToDo list
Dec 19 10:28:25 <DocScrutinizer05> now that we actually _can_ do this
Dec 19 10:28:30 <zamn900> oh god bless you
Dec 19 10:29:38 <DocScrutinizer05> jonwil: how about adding a software driven kbd-backlight ramp-up/down function into your mce?
Dec 19 10:29:52 <DocScrutinizer05> should be pretty simple
Dec 19 10:30:31 <jonwil> except that right now I cant actually compile things :)
Dec 19 10:31:16 <DocScrutinizer05> DAMN!!!
Dec 19 10:31:17 <DocScrutinizer05> ** PROBLEM Service Alert: talk.maemo.org/Disk is WARNING **
Dec 19 10:31:18 <jonwil> so its a vicious cycle, I cant do dev work so I dont have inspiration. And because I dont have inspiration, I dont have the energy to try and get e.g. a dev VM going.
Dec 19 10:31:26 <DocScrutinizer05> DISK WARNING - free space: / 9 MB (0% inode=94%):
Dec 19 10:31:40 <DocScrutinizer05> W*T*F???!!!
Dec 19 10:34:33 <jonwil> my biggest wish is that Nokia would somehow see the light and release more code. Or failing that, a copy of the latest version of IDA (and HexRays including ARM) would come down in price to the point where it was something I could buy :P
Dec 19 10:53:08 <jonwil> but yeah mostly its just the fact that it was not worth the expense to move my ancient dev box (i.e. it would have cost more to move than it was worth)
Dec 19 10:53:16 <jonwil> that and it was starting to fil
Dec 19 10:53:17 <jonwil> fail
Dec 19 11:02:51 <DocScrutinizer05> I hope you found the HDD worth backing up or moving it
Dec 19 11:03:22 <jonwil> I backed up the conrtents
Dec 19 11:03:38 <jonwil> so I have anything of value from it on an external HDD
Dec 19 11:03:49 <jonwil> just as a set of files, not an actual filesystem though :)
Dec 19 11:04:27 <jonwil> I just backed up everything over FTP that my limited-user account had read privileges for :)
Dec 19 11:05:15 <DocScrutinizer05> I would offer to find some sponsoring to get you a cheap i3 for maybe 300$, but since you say you lack motivation, it's hard to motivate donors to help you getting bored with a new box
Dec 19 11:06:37 <jonwil> yeah the hardware is not really the major issue, its the fact that I need to find actual paid work more than I need to hack Neo900 stuff
Dec 19 11:07:38 <DocScrutinizer05> I found that not having a job made me do too much maemo work. But I tried to revert that and it turned out the causality is one way
Dec 19 11:08:10 <DocScrutinizer05> IOW doing less for maemo didn't help find me a job
Dec 19 11:08:15 <DocScrutinizer05> au contraire
Dec 19 11:08:21 <DocScrutinizer05> ;-)
Dec 19 11:08:37 <DocScrutinizer05> I'm now trying to make maemo my job
Dec 19 11:09:18 <DocScrutinizer05> not that this panned out too fine so far, but at least it's a perspective
Dec 19 11:09:44 <DocScrutinizer05> oooh, didn't you say you had to move because of your job?
Dec 19 11:09:55 <jonwil> I had to move because I couldn't find a job in my old location
Dec 19 11:10:02 <DocScrutinizer05> ooh
Dec 19 11:10:06 <jonwil> that and the rent was getting too high
Dec 19 11:10:17 <DocScrutinizer05> nasty
Dec 19 11:11:19 <jonwil> Well what really happened is that the landlord told me the lease wasn't going to be renewed and after examining my prospects for finding something suitable, decided to move to another state where the availability of good rental properties is much better
Dec 19 11:11:23 <DocScrutinizer05> well, I have to get some stuff done, related to aforementioned maemo-related "job" and the whole topic. So bbl
Dec 19 11:11:51 <DocScrutinizer05> o/
Dec 19 13:50:13 <vi__> DocScrutinizer05: ping
Dec 19 14:12:47 <DocScrutinizer05> vi__: ping, 60s
Dec 19 14:13:15 <DocScrutinizer05> pong, even
Dec 19 14:13:19 <vi__> heh
Dec 19 14:13:30 <DocScrutinizer05> hurry, 20s left over
Dec 19 14:13:37 <DocScrutinizer05> ;-)
Dec 19 17:17:29 <deepy> for all you mosh naysayers
Dec 19 17:17:33 <deepy> 98 packets transmitted, 58 packets received, 40% packet loss
Dec 19 17:17:33 <deepy> round-trip min/avg/max = 236.389/656.982/3051.667 ms
Dec 19 17:18:36 <deepy> I can still irc, even with a 40 percent packetloss
Dec 19 20:02:52 <qwazix> chem|st, are you around?
Dec 19 21:02:45 <merlin1991> hbib: did anything implode yet?
Dec 19 21:05:34 <hbib> merlin1991: nope, everything works fine :)
Dec 19 23:44:52 <speps> Hi guys, n900 with a faulty eMMC here, most of the hierarchy got lost but I managed to get it mostly recovered (via ssh+rsync) without reflashing. To get it completed I need the part that is not rebuildable through packages, so here I am with the first question: /home/opt should be -> /opt, right? Thanks
Dec 19 23:45:57 <merlin1991>  yep
Dec 19 23:46:11 <merlin1991> or rather /opt is a binmount and /home/opt is the source
Dec 19 23:46:29 <merlin1991> *bindmount*
Dec 19 23:49:05 <speps> merlin1991: yeah, you're right it's not a symlink but a bind mount, thanks
Dec 19 23:51:21 <speps> merlin1991: I have an 1000:1000 /opt, probably is wrong, can you confirm it should be so or root:root?
Dec 19 23:51:54 <merlin1991> my opt is root:root
Dec 19 23:54:33 <speps> merlin1991: ok, i think i've got it, i forgot to preserve owner and group while unpacking, thanks again
Dec 19 23:58:17 <speps> merlin1991: what about /usr/share/themes? I have it as a symlink to /opt/usr/share/themes, is it right?
Dec 19 23:58:37 <merlin1991> yep
Dec 20 00:03:16 <speps> merlin1991: I'm wondering which package creates the /opt/usr/share/themes path, seems no one here
Dec 20 00:04:27 <merlin1991> maemo-optify-run-once or what it's called
Dec 20 00:04:32 <merlin1991> it moves stuff around on first boot
Dec 20 00:05:02 <merlin1991> maemo-optify-runonce is the name
Dec 20 00:05:23 <merlin1991> essentially after pr1.1 nokia realized rootfs is too small
Dec 20 00:05:36 <merlin1991> so pr1.1.1 came around which essentially was the optify stuff
Dec 20 00:05:51 <merlin1991> then 1.2 and 1.3 came which already utilized /opt
Dec 20 00:05:53 <Wizzup> interesting
Dec 20 00:06:14 <merlin1991> I think they added Qt with 1.1 or 1.2 so a lot more space was needed
Dec 20 00:06:51 <merlin1991> Qt libs in /opt/libs alone have 43.6 mb
Dec 20 00:07:22 <merlin1991> and all user apps used to live on rootfs aswell ...
Dec 20 00:08:20 <speps> merlin1991: so probably the whole /opt/usr is created via that script, right?
Dec 20 00:08:27 <merlin1991> yep
Dec 20 00:09:09 <merlin1991> check the scripts in /usr/sbin/maemo-optify*
Dec 20 00:09:31 <speps> merlin1991: could you ls -R /opt/usr > somefile and pastebin it, so I can check by myself how it should be?
Dec 20 00:10:04 <merlin1991> sure
Dec 20 00:10:20 <Wizzup> Dont you think reflashing is easier?
Dec 20 00:11:54 <speps> Wizzup: if I do that again I could puke ;) now it is more a matter of principle :)
Dec 20 00:12:53 <merlin1991> speps: http://cdnm.at/~christian/maemo/usr.txt
Dec 20 00:13:14 <Wizzup>  speps I recall it being better to flash...
Dec 20 00:13:18 <Wizzup> but ok :)
Dec 20 00:14:01 <speps> Wizzup: long story short, never been in maemo world, a week ago a guy sold me this n900 for few euros, but I discovered the internal eMMC is partially damaged and unfortunately the first sectors, so now I finding a setup that could work fine
Dec 20 00:14:27 <Wizzup> speps: if you reflash it should get rid of bad sectors
Dec 20 00:14:29 <Wizzup> or rather, not use them
Dec 20 00:14:30 <Wizzup> AFAIK
Dec 20 00:16:10 <speps> Wizzup: tried a lot of times, unfortunately having the first sectors damaged means that the partition table gets randomly damaged so it's not a metter of reflashing
Dec 20 00:16:24 <speps> merlin1991: thanks m8, checking right now
Dec 20 00:16:32 <Wizzup> weird, the card should take care of that...
Dec 20 00:18:03 <speps> Wizzup: it wouldn't be damaged so ;) or maybe is not the card itself, but I'm not an hardware guy, so I'm still facing it
Dec 20 00:28:39 <speps> merlin1991: I wonder why the /opt/usr structure is not handled by a package instead of script, maybe a backward compatibility reason? In case of conflicts it could be handled in a pre-install script too
Dec 20 00:29:42 <merlin1991> because /opt/* got created after the fact
Dec 20 00:30:16 <merlin1991> but it couldn't get created by new packages because that would have filled up rootfs when extracting / dl all the packages that needed to move content around
Dec 20 00:38:37 <speps> merlin1991: I admit I'm really new to n900, but on a first look the script looks to check whenever there is enough space to do the operation, so it could be potentially used in a pre-install script of a new packages, when and who calls the script?
Dec 20 00:53:28 <speps> merlin1991: sorry I gave you the wrong command, I meant ls -lR /opt/usr (it wouldn't show eventual symlinks if it has)
Dec 20 00:54:15 <merlin1991> I noticed that and made a new txt, but I think I did that before posting it, anyways try F5 :)
Dec 20 00:59:40 <speps> merlin1991: the file is still the same, but I don't it depends on the fact that I'm more an Ctrl+R over F5 guy ;) btw it could be indifferent, probably /opt/usr has not symlinks, you can confirm
Dec 20 01:50:10 <int_ua> Anyone knows the best way to push two arguments through g_idle_add() ?
**** ENDING LOGGING AT Fri Dec 20 03:00:00 2013