**** BEGIN LOGGING AT Mon Jan 16 03:00:01 2017
Jan 16 09:57:15 <KotCzarny> hmm, idea. it shouldnt be hard to replace broken search on t.m.o with google widget
Jan 16 09:57:49 <KotCzarny> who is tmo admin? xes?
Jan 16 09:59:52 <warfare> chem|st is. 
Jan 16 10:00:18 <bencoh> google widget? are you serious? ...
Jan 16 10:08:18 <KotCzarny> bencoh, most people would use google to search tmo anyway
Jan 16 10:08:56 <KotCzarny> or even better it could be a simple form to redirect to google, ie google.com?search?q=something
Jan 16 10:09:04 <KotCzarny> so not even a widget
Jan 16 10:13:23 <KotCzarny> bencoh, unless you know how to make tmo search work (i suspect some indexing service is not running)
Jan 16 10:34:59 <chem|st> KotCzarny: broken search widget?! powersearch works, that is google search
Jan 16 10:36:39 <KotCzarny> wha'ts powersearch?
Jan 16 10:37:01 <chem|st> http://talk.maemo.org/msearch.php
Jan 16 10:37:16 <KotCzarny> um, but the thing on the top right doesnt work
Jan 16 10:37:23 <KotCzarny> maybe you should update this part?
Jan 16 10:37:24 <chem|st> KotCzarny: it is in the menu the first item...
Jan 16 10:37:32 <chem|st> KotCzarny: update what?
Jan 16 10:37:53 <KotCzarny> on the top right corner you have search form input
Jan 16 10:38:02 <KotCzarny> which is broken (and uses search.php)
Jan 16 10:38:12 <KotCzarny> replace it to use msearch then?
Jan 16 10:38:48 <chem|st> this would force people to use google
Jan 16 10:38:53 <KotCzarny> nono
Jan 16 10:39:07 <KotCzarny> if msearch works, then replace top right search to use msearch
Jan 16 10:39:31 <chem|st> KotCzarny: msearch is google
Jan 16 10:40:09 <KotCzarny> what i meant originally is that topright search box with orange 'go' button is broken
Jan 16 10:40:41 <chem|st> in what way broken, the returns are odd sometimes but in general it works
Jan 16 10:40:53 <KotCzarny> try searching for usb
Jan 16 10:40:54 <KotCzarny> :)
Jan 16 10:41:17 <chem|st> that is if you get more than 1k returns
Jan 16 10:41:26 <KotCzarny> fix the error msg then?
Jan 16 10:41:42 <chem|st> it should actually stopp at 500 but for some reason does not
Jan 16 10:41:43 <KotCzarny> right now it is completely misleading
Jan 16 10:41:58 <KotCzarny>     Sorry - no matches. Please try some different terms. 
Jan 16 10:42:25 <KotCzarny> searching for 'usb fix' also failed
Jan 16 10:42:40 <chem|st> I know but I cannot change a reply that is actually ok just because a bug in vB does that 
Jan 16 10:42:55 <KotCzarny> and if you require 4 chars or more, maybe add such information too?
Jan 16 10:43:35 <chem|st> I tried to fix it on several occasions... sometimes I then get a reply working again and after a while it stopps working, so it has to do with the cache
Jan 16 10:43:41 <chem|st> or call it index
Jan 16 10:43:59 <chem|st> reindexing is 72h downtime of the forum
Jan 16 10:44:39 <KotCzarny> my suggestion is to add information on the search.php results page that would explain that searching requires 4chars or more and that it can fail when too many results are returned
Jan 16 10:44:58 <chem|st> or well, better said to be 7 days reindexing with no forum downtime, that is what it took a couple of years ago
Jan 16 10:46:05 <KotCzarny> do it semi-annually then with sticky warning ?
Jan 16 10:46:58 <chem|st> I can revert the search function to keywords instead of fulltext
Jan 16 10:47:26 <chem|st> might work better in terms of you get returns but less in terms of quality
Jan 16 10:47:41 <KotCzarny> maybe make it use keywords when fulltext fails?
Jan 16 10:48:45 <bencoh> waitamin ... search fails when there are too many results?
Jan 16 10:49:16 <KotCzarny> :>
Jan 16 10:49:18 <Michael_a380> I don't understand much of this, but how about a 'permenant / static messege' about the search limitations & possible 'search errors' ?
Jan 16 10:49:23 <KotCzarny> also on 3 letter words
Jan 16 10:50:31 <chem|st> KotCzarny: the minimum search text is set to 3... something is wrong with that indexer
Jan 16 10:52:34 <chem|st> " MySQL Fulltext has its own minimum word length as well that must be changed at the server level." I might have found the issue
Jan 16 10:55:24 <chem|st> setting it to 4 still does not tell that "usb" is too short
Jan 16 11:05:26 <chem|st> MySQL fulltext min word length is defaulting to 4
Jan 16 11:07:20 <bencoh> huhu
Jan 16 11:07:25 <bencoh> well ... :)
Jan 16 12:42:50 <sicelo-> stuck
Jan 16 12:43:22 <sicelo-> sorry about that one
Jan 16 13:38:35 <NeKit> how different Fremantle UI inside Scratchbox target is compared to real device?
Jan 16 15:31:54 <Vajb> is it possible to lock n900 volume keys while screen is locked?
Jan 16 15:36:42 <DocScrutinizer05> err yes?
Jan 16 15:39:40 <DocScrutinizer05> at least supposed to, though I#m not all that sure this actually _works_: etc/mce/mce.ini http://paste.opensuse.org/27541461
Jan 16 15:40:51 <DocScrutinizer05> given that vol+/- are actually mapped to F7/9 or somesuch, this is exactly the dial tu tweak for locking volume keys
Jan 16 15:43:07 <DocScrutinizer05> freemangordon: ^^^ doesn't work. Is this a mce regression in CSSU?
Jan 16 15:43:26 <DocScrutinizer05> freemangordon: how would I even check which version of mce I'm running?
Jan 16 15:44:08 <Vajb> it is set to 1 already, hmm
Jan 16 15:44:29 <DocScrutinizer05> yes, here too
Jan 16 15:44:41 <DocScrutinizer05> what system you're running?
Jan 16 15:45:12 <Vajb> cssu-testing
Jan 16 15:45:31 <Vajb> and volume keys r working while screen is locked
Jan 16 15:45:59 <Vajb> and that is kinda inconvenient if phone is in my pocket and starts to adjust volume to max...
Jan 16 15:46:01 <DocScrutinizer05> same here
Jan 16 15:46:57 <Vajb> so no cookie yet. I'll have to contiue work, but i'll read backlog if u come up with something
Jan 16 15:47:00 <DocScrutinizer05> IroN900:~# ls -l /sbin/mce; md5sum /sbin/mce
Jan 16 15:47:01 <DocScrutinizer05> -rwxr-xr-x 1 root root 81160 2015-04-05 16:56 /sbin/mce
Jan 16 15:47:03 <DocScrutinizer05> 1edd10407336c736c0466bbcfe2084c9  /sbin/mce
Jan 16 15:52:09 <DocScrutinizer05> I got a /home/opt/var/cache/apt/archives/mce_1.8.126+0m5_armel.deb with -rwxr-xr-x root/root     80760 2010-07-29 14:30 ./sbin/mce
Jan 16 15:52:26 <DocScrutinizer05> ~pkg
Jan 16 15:52:31 <infobot> methinks pkg is http://maemo.org/packages/
Jan 16 15:52:32 <DocScrutinizer05> ~pkgs
Jan 16 15:52:32 <infobot> #maemo pkg is probably http://maemo.org/packages/
Jan 16 15:54:59 <DocScrutinizer05> http://wiki.maemo.org/Community_SSU/Changelog Tmaemo11 >>mce  * FOSS release  * Fixes for upstream kernel<<
Jan 16 15:56:04 <DocScrutinizer05> freemangordon: wasn't it you who made FOSS mce? please check etc/mce/mce.ini http://paste.opensuse.org/27541461  vs http://maemo.org/packages/package_instance/view/fremantle_ssu_pr1-2_armel/mce/1.8.118+0m5/  -- REGRESSION?
Jan 16 15:58:18 <DocScrutinizer05> http://paste.opensuse.org/28099749
Jan 16 15:59:03 <DocScrutinizer05> Vajb: would you like to install http://maemo.org/packages/package_instance/view/fremantle_ssu_pr1-2_armel/mce/1.8.118+0m5/ ? might help
Jan 16 16:00:11 <DocScrutinizer05> oops sorry, nevermind
Jan 16 16:00:26 <DocScrutinizer05> http://maemo.org/packages/package_instance/view/fremantle_root_pr1-2_armel/mce/1.8.126+0m5/
Jan 16 16:02:52 <DocScrutinizer05> http://paste.opensuse.org/9778156
Jan 16 16:02:57 <DocScrutinizer05> Vajb: ^^^
Jan 16 16:14:55 <sicelo-> nnot to belittle your problem Vajb ... my volume buttons need 'intention' to press, and they aren't so soft to press too ... can't imagine what would activate them in pocket
Jan 16 16:18:14 <sicelo-> that 'thing' sounds likely to even activate power button and turn off/on the device
Jan 16 16:19:45 <freemangordon> DocScrutinizer05: what is the regression?
Jan 16 16:19:47 <DocScrutinizer51> nope, I just installed mce*126 and  it nuked community-PR but didn't fix vol kets
Jan 16 16:20:23 <freemangordon> yes, this change (enable vol keys even when device is locked) was introduced in pr1.2 iirc
Jan 16 16:20:36 <DocScrutinizer51> freemangordon: vol keysl not locked
Jan 16 16:20:49 <freemangordon> yes, this is intentional
Jan 16 16:20:58 <DocScrutinizer51> FFS rhere's a config option for that
Jan 16 16:21:05 <freemangordon> and?
Jan 16 16:21:21 <DocScrutinizer51> it doesn't do a thing
Jan 16 16:21:41 <freemangordon> does it do a thing on pr1.2 mce?
Jan 16 16:21:47 <DocScrutinizer51> no
Jan 16 16:21:53 <DocScrutinizer51> see aboce
Jan 16 16:21:59 <freemangordon> so? what is regressed?
Jan 16 16:22:00 <DocScrutinizer51> above
Jan 16 16:22:09 <freemangordon> yep, saw it
Jan 16 16:22:18 <DocScrutinizer51> none obviously, see above
Jan 16 16:22:46 <DocScrutinizer51> the regreesion is with pr12 mce
Jan 16 16:23:00 <DocScrutinizer51> which fucked up
Jan 16 16:23:38 <freemangordon> ah, you mean that because they ignore the settings file, there is a bug? sounds same
Jan 16 16:23:44 <freemangordon> *sane
Jan 16 16:25:07 <freemangordon> DocScrutinizer05: does cssu mce segfaults as well?
Jan 16 16:25:10 <DocScrutinizer51> yep
Jan 16 16:25:24 <freemangordon> but, there is another mce instance running, right?
Jan 16 16:25:27 <DocScrutinizer51> yes
Jan 16 16:25:36 <DocScrutinizer51> see pastebin
Jan 16 16:26:10 <freemangordon> the one with the segfault?
Jan 16 16:26:17 <DocScrutinizer51> it has version and segf
Jan 16 16:27:03 <freemangordon> hmm. DocScrutinizer05please open an issue on github
Jan 16 16:29:55 <freemangordon> wtf?
Jan 16 16:30:50 <freemangordon> DocScrutinizer05: https://github.com/community-ssu/mce/issues
Jan 16 16:34:17 <KotCzarny> no issue, perfect package
Jan 16 16:34:21 <KotCzarny> :)
Jan 16 16:34:39 <DocScrutinizer51> k, laqyer. roaming
Jan 16 16:34:50 <freemangordon> KotCzarny: the wtf ^^^ was because issues were not enabled for this package :D
Jan 16 16:34:55 <DocScrutinizer51> kater
Jan 16 16:34:58 <KotCzarny> ;)
Jan 16 16:34:59 <freemangordon> ok
Jan 16 16:35:00 <DocScrutinizer51> dang
Jan 16 16:35:06 <freemangordon> later
Jan 16 16:35:07 <KotCzarny> no issues, no problem ;)
Jan 16 16:35:08 <freemangordon> :)
Jan 16 16:35:14 <freemangordon> yep
Jan 16 16:36:23 <DocScrutinizer51> Ill check if 126 also segfs
Jan 16 16:37:25 <DocScrutinizer51> yep, all the  same
Jan 16 16:38:27 <DocScrutinizer51> afk
Jan 16 16:39:51 <Vajb> hmm so no need to install anything new since it wont change a thing?
Jan 16 16:40:56 <Vajb> DocScrutinizer05: i already have newer mce
Jan 16 16:41:22 <Vajb> mce: Installed: 1.8.127.3+0m5
Jan 16 16:47:09 <Vajb> sicelo-: isn't powerkey locked when device is?
Jan 16 16:47:28 <bencoh> huh?
Jan 16 16:47:49 <bencoh> no it isn't :)
Jan 16 16:48:12 <Vajb> oh dang
Jan 16 16:48:12 <KotCzarny> hum, it isnt?
Jan 16 16:48:24 <KotCzarny> i've thought all keys are locked (- powerkey)
Jan 16 16:48:26 <bencoh> funny how I almost forgot about the lock switch
Jan 16 16:48:32 <bencoh> I just never use it
Jan 16 16:48:35 <bencoh> (at all)
Jan 16 16:48:37 <Vajb> i hope it wont turn my device of then..
Jan 16 16:49:47 <sicelo-> yes it's not :-)
Jan 16 16:55:35 <DocScrutinizer05> FYI: http://paste.opensuse.org/3657714
Jan 16 17:00:30 <DocScrutinizer05> Vajb: I thought DOWNgrading from 127 to 126 would solve the problem, but the regression sneaked in somewhere around 115
Jan 16 17:01:35 <DocScrutinizer05> bencoh: like I almost never use power button
Jan 16 17:01:56 <DocScrutinizer05> except for checking time of day
Jan 16 17:02:25 <bencoh> same :)
Jan 16 17:02:49 <bencoh> well, or power device off with systemui menu
Jan 16 17:03:18 <bencoh> I use a wayv gesture to lock phone
Jan 16 17:06:29 <DocScrutinizer05> oh!
Jan 16 17:06:44 <DocScrutinizer05> now that's geeky :-)
Jan 16 17:08:27 <bencoh> :>
Jan 16 17:08:31 <DocScrutinizer05> given the extremel<y pathetic accelerometer integration, isn't it power greedy to poll the accel to detect gestures like that?
Jan 16 17:08:46 <bencoh> touchscreen gestures
Jan 16 17:08:49 <DocScrutinizer05> oh
Jan 16 17:09:45 <bencoh> I patched wayv a bit to adapt it to maemo/n900, and configured it to only capture mouse in a 1px (or 2px)-large area on the side of the screen
Jan 16 17:09:47 <DocScrutinizer05> for accel, Nokia failed epically to invent a decent API
Jan 16 17:10:08 <bencoh> meaning that all my wayv gestures start from the side of the screen
Jan 16 17:10:22 <bencoh> but it can continue from there all screen over
Jan 16 17:11:01 <bencoh> so I have a few of them to switch to desktop, task launcher, app launcher, or lock device
Jan 16 17:11:36 <DocScrutinizer05> ((accel)) the chip's filter and trigger settings are HARDCODED into driver, go figure!
Jan 16 17:12:32 <DocScrutinizer05> and if that wasn't already bad enough, they even used pretty bad and silly settings
Jan 16 17:12:38 <bencoh> filter and trigger ... is it supposed to be done in hw by the chip itself?
Jan 16 17:12:46 <DocScrutinizer05> yes
Jan 16 17:12:47 <bencoh> I know some TI chips have support for tiny firmwares
Jan 16 17:12:52 <bencoh> oh, nice then
Jan 16 17:13:07 <bencoh> and you can configure it to detect free fall, or "tap"/double-tap
Jan 16 17:14:04 <DocScrutinizer05> you can set a min and max threshold on 2 of the 3 axes. and you can apply a highpass filter to only fast enough movement from any position into any other position would trigger an IRQ. 
Jan 16 17:14:28 <DocScrutinizer05> Nokia set a "DC" threshold on just one axis iirc
Jan 16 17:15:07 <DocScrutinizer05> it was just as much as they needed to detect "screen down"
Jan 16 17:15:31 <DocScrutinizer05> pathetic
Jan 16 17:17:05 <DocScrutinizer05> what's worse: this is in kernel driver, and API is fubar so there's not even a possible way to fix this in userland
Jan 16 17:18:59 <DocScrutinizer05> just like they fucked up LED pattern engine LS5523, they don't use all the memory this chip offers, and they don't expose the var registers (or any other registers on that topic)
Jan 16 17:19:41 <DocScrutinizer05> at least that API isn't *totally* fubar, it could get augmented and fixed
Jan 16 17:29:48 <bencoh> well, overengineering it would have been easy as well, so ...
Jan 16 17:29:53 <bencoh> can't blame them
Jan 16 17:31:47 <Vajb> KotCzarny: how is that imgur coming together?
Jan 16 17:39:23 <KotCzarny> vajb: i was preparing opensource oscp release, will try looking into imgur plugin tomorrow
Jan 16 17:41:32 <bencoh> oh, really? :)
Jan 16 17:41:40 <bencoh> nice
Jan 16 17:56:29 <Vajb> cool :p
Jan 16 20:59:09 <NeKit> how much the system inside SDK is different from real device?
Jan 16 21:20:26 <DocScrutinizer05> it lacks quite some sensors, obviously
Jan 16 21:20:49 <DocScrutinizer05> and iirc it runs x86 instead arm
Jan 16 21:21:21 <DocScrutinizer05> also no phone and prolly no audio either
Jan 16 22:54:22 <DocScrutinizer05> how the heck would I disable the twl4030 keypad IRQ?  sth about:
Jan 16 22:54:24 <DocScrutinizer05> --w------- 1 root root 4096 2017-01-16 23:52 /sys/bus/platform/drivers/twl4030_keypad/bind
Jan 16 22:54:25 <DocScrutinizer05> --w------- 1 root root 4096 2017-01-16 23:52 /sys/bus/platform/drivers/twl4030_keypad/unbind
Jan 16 22:54:27 <DocScrutinizer05> ?
Jan 16 23:00:51 <DocScrutinizer05> Vajb: to lock kbd completely: ## echo 'twl4030_keypad' >/sys/bus/platform/drivers/twl4030_keypad/unbind; ##  to re-enable: ## echo 'twl4030_keypad' >/sys/bus/platform/drivers/twl4030_keypad/bind; ##   
Jan 16 23:01:10 <DocScrutinizer05> swiftly tested, seems to work with no obvious negative side effects
Jan 16 23:02:12 <DocScrutinizer05> Vajb: now what you need is a way to make that happen on device/screen lock and unlock respectively
Jan 16 23:03:03 <DocScrutinizer05> freemangordon: ^^^ idealy this becomes a patch to mce, right?
Jan 16 23:04:43 <DocScrutinizer05> Vajb: until then, I'm pretty sure you can clutch together sth with dbus-scripting, triggering the according commands as mentioned above on signals that indicate tklock / unlock
Jan 16 23:06:40 <DocScrutinizer05> I suggest:
Jan 16 23:06:42 <DocScrutinizer05> signal sender=:1.14 -> dest=(null destination) serial=525 path=/com/nokia/mce/signal; interface=com.nokia.mce.signal; member=tklock_mode_ind
Jan 16 23:06:43 <DocScrutinizer05>    string "locked"
Jan 16 23:07:30 <DocScrutinizer05> aka "mce sends signal: 'device locked' "
Jan 16 23:08:18 <DocScrutinizer05> see pkg "dbus-scripting"
Jan 16 23:13:01 <DocScrutinizer05> http://paste.opensuse.org/74779136
Jan 16 23:13:36 <DocScrutinizer05> note that I strongly recommend to use dbus-scripting instead dbus-monitor
Jan 16 23:43:34 <DocScrutinizer05> not recommended (but works):   dbus-monitor --system|grep --line-buffered -A1 'interface=com.nokia.mce.signal; member=tklock_mode_ind'|while read; do case $REPLY in *string\ \"unlocked\"*) echo UNLOCKED; echo 'twl4030_keypad' >/sys/bus/platform/drivers/twl4030_keypad/bind;; *string\ \"locked\"*) echo LOCKED; echo 'twl4030_keypad' >/sys/bus/platform/drivers/twl4030_keypad/unbind;; esac; done
Jan 16 23:50:30 <DocScrutinizer05> no idea if maybe (from syslog) this would be better suited:  Jan 17 00:49:21 IroN900 systemui-tklock[1085]: Method call received from: :1.14, iface: com.nokia.system_ui.request, method: tklock_close
Jan 17 00:43:19 <jerardepardieu> ~dnsmasq
**** ENDING LOGGING AT Tue Jan 17 03:00:02 2017