IRC log of #maemo-ssu for Tuesday, 2012-11-06

*** luf has joined #maemo-ssu00:01
*** MrPingu has quit IRC00:28
*** MrPingu has joined #maemo-ssu00:29
*** deepy\SIGSEGV has quit IRC00:45
*** deepy\SIGSEGV has joined #maemo-ssu00:46
*** luf has quit IRC00:49
*** MrPingu has quit IRC00:58
*** infobot has quit IRC01:15
*** infobot has joined #maemo-ssu01:17
*** ChanServ sets mode: +v infobot01:17
*** MrPingu has joined #maemo-ssu01:17
*** arcean_ has quit IRC01:19
*** arcean_ has joined #maemo-ssu01:30
*** ZogG has quit IRC01:30
*** andre__ has quit IRC01:43
*** ZogG has joined #maemo-ssu01:45
*** andre__ has joined #maemo-ssu01:59
*** arcean_ has quit IRC03:38
*** guly has quit IRC04:12
*** guly has joined #maemo-ssu04:37
*** M4rtinK has quit IRC04:50
*** nox- has quit IRC05:03
*** amiconn_ has joined #maemo-ssu05:40
*** amiconn has quit IRC05:40
*** amiconn_ is now known as amiconn05:40
*** DocScrutinizer05 has quit IRC06:04
*** DocScrutinizer05 has joined #maemo-ssu06:04
*** Pali has quit IRC06:12
*** Jade has quit IRC06:32
*** Jade has joined #maemo-ssu06:32
*** Jade has joined #maemo-ssu06:32
*** Jade has quit IRC06:33
*** Jade has joined #maemo-ssu06:35
*** Jade has joined #maemo-ssu06:35
*** Jade has quit IRC06:39
*** Jade has joined #maemo-ssu06:40
*** Jade has joined #maemo-ssu06:40
*** Jade has quit IRC06:40
*** Jade has joined #maemo-ssu06:53
*** Jade has joined #maemo-ssu06:53
*** povbot has joined #maemo-ssu08:41
*** ChanServ sets mode: +v povbot08:41
*** ruskie has quit IRC08:42
*** ruskie has joined #maemo-ssu09:01
*** luf has joined #maemo-ssu09:35
*** dhbiker has quit IRC09:59
*** dhbiker has joined #maemo-ssu09:59
*** X-Fade has quit IRC11:02
*** X-Fade has joined #maemo-ssu11:02
lufDocScrutinizer51: no damage for N900 or carkit (unless angry user don't destroy it) :D So no regression :D12:05
*** dhbiker has quit IRC12:26
*** MrPingu has quit IRC12:35
*** merlin1991 has quit IRC13:01
*** merlin1991 has joined #maemo-ssu13:01
*** arcean has joined #maemo-ssu13:07
*** M4rtinK has joined #maemo-ssu13:17
*** ivgalvez has joined #maemo-ssu13:19
*** Pali has joined #maemo-ssu14:11
*** dhbiker has joined #maemo-ssu14:15
*** krayon has quit IRC14:24
*** lizardo has joined #maemo-ssu14:46
DocScrutinizer05luf: OK :-)14:46
*** arcean has quit IRC15:06
*** ivgalvez has quit IRC15:32
merlin1991luf: do we know if the patch is good by now?15:54
*** Woody14619a has joined #maemo-ssu15:55
*** Woody14619a has joined #maemo-ssu15:55
lufmerlin1991: The patch seems to be ok now.15:55
lufWhen you want to release new -testing?15:55
*** ivgalvez has joined #maemo-ssu15:57
*** Woody14619 has quit IRC15:58
*** dafox has joined #maemo-ssu16:01
*** arcean has joined #maemo-ssu16:08
*** arcean has quit IRC16:17
*** krayon has joined #maemo-ssu16:20
merlin1991as soon as I can :)16:30
lufOps :)16:41
lufI hope I'll push the git today or tomorrow with the obexd fix.16:42
*** dafox has quit IRC16:53
chem|stmerlin1991: as soon as I am done building that stuff I will shoot you a list to checkback on my doings, I am not really sure it is all right what I do atm...16:56
merlin1991take care not to take the new testing stuff ;)16:57
chem|stmerlin1991: currently I am at stuff some month old apart from curl17:09
merlin1991:)17:09
chem|stqt-x11 takes ages to build17:09
merlin1991yep17:09
merlin1991openssl is almost as bad17:09
chem|stlove it!17:09
merlin1991though in the case of openssl it's a billion test that are run, whils qith qt it's the friggin html docs17:09
chem|st*sigh*17:10
merlin1991s/whils qith/whilst with/17:11
infobotmerlin1991 meant: though in the case of openssl it's a billion test that are run, whilst with qt it's the friggin html docs17:11
freemangordonmerlin1991: hmm, iirc it is webkit js that takes ages17:11
merlin1991I was up way too long today having fun with kvm, when I finally started to pester people in #virt the my typo to proper words ratio was beyond readable :D17:12
chem|st^^17:12
merlin1991freemangordon:  possibly webkit aswell, last time I built qt it was the harm qt with meegotouch stuff, back then it was the docs that took forver :D17:15
merlin1991*forever*17:15
merlin1991btw I have to say qemu-kvm from debian-testing together with virt-manager works quite well to virtualize win stuff (once one removed the shortcomings of the default windows template and installs the virtio drivers from readhat into the guest)17:17
chem|stmerlin1991: for the record it took ages for html stuff now it takes ages for webikit js...17:23
chem|stI think those ssl test are not even close as boring...17:24
*** dafox has joined #maemo-ssu17:37
*** arcean has joined #maemo-ssu17:55
chem|stmerlin1991: modest? commit 17e9406 or latest?18:14
*** Jade has quit IRC18:15
*** Jade has joined #maemo-ssu18:15
*** Jade has quit IRC18:15
*** Jade has joined #maemo-ssu18:15
chem|stI guess the later is not in last testing is it?!18:16
chem|stgot it18:17
*** freemangordon has quit IRC18:17
*** lizardo has quit IRC18:17
*** luf has quit IRC18:17
*** freemangordon has joined #maemo-ssu18:17
merlin1991chem|st: you don't need modest at all18:18
merlin1991ah wait you to18:18
merlin1991*do*18:18
freemangordonyes, he wants it ;)18:18
freemangordonchem|st: aiui you should pull latest master in stable, merlin1991, agree?18:19
merlin1991freemangordon: yep sounds about right18:20
freemangordontinymail as well18:20
merlin1991tinymail is later in the alphabet ;)18:21
freemangordonaah, ok :D18:21
merlin1991btw chem|st don't take obexd18:21
merlin1991it's faulty18:22
chem|sthmm just build the one from testing18:24
chem|stmodest*18:24
chem|stnot obex18:24
chem|stI am reading the chan18:24
kerioi thought obexd was fixed-ish18:25
keriooh, not pushed yet18:25
chem|stwierd I get a all up to date with git merge master... I am doing something wrong here...18:26
chem|stlatest in my log is palis commit not the newer by fmg18:27
*** ivgalvez has quit IRC18:30
chem|stanyone?18:30
chem|stI merged it now with the commit number but is that the right way?18:30
chem|stwonderfull now I am back where I was and it does not built anymore &"$%@18:36
chem|staclocal: macro `gl_FUNC_ARGZ' required but not defined18:38
chem|stmake: *** [configure-stamp] Error 118:38
chem|stfor god sake I just built libxml2 in the mean time18:42
chem|stso that is broken for me now...18:43
chem|stany ideas?18:43
chem|stI guess something pulled in another version of autoconf or something when I installed python dev stuff for xml218:44
*** NIN101 has joined #maemo-ssu18:44
*** Milhouse has quit IRC19:03
*** Milhouse has joined #maemo-ssu19:07
*** arcean has quit IRC19:08
*** arcean has joined #maemo-ssu19:12
merlin1991chem|st: a merge is local only19:25
merlin1991you need to git fetch changes from the repos first to get everything up to date :)19:26
Palimerlin1991, I have new version of battery status menu plugin19:27
Palicompatible with BME and with kernel drivers19:27
Palibased on MAG version19:27
merlin1991psuh19:27
merlin1991*push*19:27
Paliwhat do you think can it go to CSSU?19:27
PaliI will push it into my git branch19:28
merlin1991okay, well do we know if the default status menu plugin does anything besides displaying?19:28
Paliit play souns & show hildon banners19:28
Palibattery is empty19:28
Palibattery is full19:29
Palicharging19:29
*** arcean has quit IRC19:29
PaliI looked into nokia status menu battery binary, exported function calls and there is nothing more I think19:29
merlin1991does yours do all the funky stuff aswell?19:30
merlin1991(sound and banners)19:30
Palicode is here: https://gitorious.org/~pali/community-ssu/pali-status-area-applet-battery19:30
Paliyes19:30
Palithere is only one small problem: missing space between icon and text19:30
Paliso layout is not same19:31
freemangordonPali, merlin1991: Seems I am stupid today, look at that19:31
freemangordonalarms_events[i++] = alarms_events[i]19:31
freemangordonwill that do what I think it will, i.e. alarms_events[i] = alarms_events[i+1]19:31
merlin1991am i right saying thats arr[i] = arr[i]; i++; ?19:31
Palimerlin1991, read commit message: https://gitorious.org/~pali/community-ssu/pali-status-area-applet-battery/commit/1393753876927073058b5d32c917c0e37a8dc3fb19:32
freemangordonmerlin1991: hmm, no19:32
freemangordonthe question is if it is ^^^ or alarms_events[i+1] = alarms_events[i+1]19:32
merlin1991hm isn't the right side evaluated first19:32
PaliI think that specification did not specify order when i++ would be increased19:32
freemangordonafaik no19:32
freemangordonI think it specifies, that is why there is a difference between i++ and ++i19:33
merlin1991no ++i is increment before evaluation19:33
freemangordonexactly19:33
merlin1991but I'm more wondering if it evaluets left or right from = first19:34
freemangordondeffinitely it is left first19:34
kerioit's definetely unspecified19:34
freemangordon(afaik)19:34
Palibut I think it is not specified if i++ will be increased after (alarms_events[i++]) or after (alarms_events[i++] = alarms_events[i])19:34
merlin1991hm then you should have arr[i] = arr[i+1]19:34
freemangordonkerio: since when you know C?19:34
kerioyou're changing and accessing the same variable without a sequence point inbetween19:34
freemangordonmerlin1991: yeah, I was trying to do it in the unreadable way :D19:35
freemangordonkerio: defien sequence point19:35
merlin1991tbh don't count on me being right I could say totall bs atm :D19:35
freemangordonwell, will do it in the old-fashined and readable way :D:D:D19:35
keriofreemangordon: http://c-faq.com/expr/seqpoints.html19:36
keriofrom the comp.lang.c FAQ19:36
kerioalso, since when *you* know C?19:36
freemangordonkerio: never heard of that term, but if we use it, there IS seq point, as you have to index the array19:37
freemangordonkerio: for the last 20 or so years19:37
freemangordonany other questions?19:37
freemangordonhmm, make them 2519:37
kerioand you've never read anything about undefined things in the C standard?19:37
freemangordonI forgot how old I am :D:D:D19:38
freemangordonnever19:38
freemangordon:P19:38
merlin1991dafuq how old are you freemangordon?19:39
freemangordon37 :(19:39
freemangordonwell, not 25, 21-22 then19:40
ShadowJKUndefined things are things that one compiler might do something consistent for, over and over again, but another compiler might do something else consistently, and a third compiler might do something random each time :P19:40
ShadowJK(and there was that one compiler that wiped your harddrive)19:40
merlin1991anybody in here who isn't old in comparison to me?19:41
kerioif 91 is your birthyear, i'm younger than you19:42
merlin1991yay I'm not the youngest :)19:42
merlin1991we kinda have an upside down structure here, the (almost)youngest dude sits on top19:42
freemangordonShadowJK: I know what you mean, but if we take C before the modern era with all that extensions and stuff, there was no such thing like "undefined" behaviour in language itself19:42
freemangordonjust undefined compiler behaviour19:42
kerioANSI C specifies a lot of things as undefined19:42
*** arcean has joined #maemo-ssu19:43
kerioliterally undefined, as in "this thing is undefined"19:43
freemangordonother term for "compiler specific implementation"19:43
kerioyes, the correct term19:43
freemangordonthough last time I check ++,--, etc was pretty defined in terms of evaluation. I am just not dure if gcc is aware of that :D19:44
freemangordon*checked19:44
freemangordonthe fuck. *sure19:44
freemangordonanyway, I am back to stupid alarm dialog19:45
kerioexcept that *they are not*, if you're accessing the value of the variable you incremented somewhere else in the same expression19:45
ShadowJKWell your example is good example of undefined behaviour? Compiler is free to evaluate it in whatever order it wants?19:45
keriofor instance, i = i++19:45
freemangordon^^ makes no sense, a better example is *p++ = *p19:46
kerioShadowJK: the compiler is free to show goatse on the screen and then rm -rf19:46
freemangordonand it IS defined19:46
ShadowJKhttp://stackoverflow.com/questions/11657100/is-this-undefined-behavior-or-implementation-defined :D19:47
freemangordonwe were talking about C, case closed AIUI :D19:49
ShadowJKBut he also quoted C :)19:50
freemangordonyeah, but my englis is not good enough to understand that cryptic sentence :D19:50
freemangordon*english19:50
freemangordonand still, this is C99 which is quoted19:51
*** dhbiker has quit IRC19:56
*** arcean has quit IRC19:58
keriofreemangordon: hold on, *p++ = *p is just *p++, isn't it19:58
ShadowJKin any case, I would expect your original expression to be equivalent of: alarms_events[i] = alarms_events[i]; i = i + 1;  ?19:59
freemangordonkerio: i think so20:00
freemangordonShadowJK: why? we need first [] evaluated20:00
freemangordoni.e. as [i], ++ applied and then second [] evaluated20:01
ShadowJKI'd expect the ++ to be executed after the ;20:01
freemangordonwhy?20:01
kerioShadowJK: it's just guaranteed to have been executed once you reach ;20:02
freemangordonhow is that different to *p++20:02
ShadowJKBut now that I google it, I get the impression that it's undefined and you might get either result, depending on the compiler20:02
ShadowJKor no result at all20:02
kerioand there's no sequence point in [] or = anyway, so it's probably a moot point20:02
ShadowJKwith -Wall, gcc actually gives a warning both with -std=c89 and -std=c99 about your example too20:05
ShadowJKAlso with gnu89 and gnu9920:05
freemangordonwell, it could be compiler depenedent. I'll try m$ C and gcc wneh I have some free time, to see what will be the output20:05
freemangordonwhat waring?20:05
freemangordon*warning20:05
ShadowJKoperation on ā€˜iā€™ may be undefined20:05
freemangordon"undefined..."20:05
freemangordonok20:05
ShadowJKWhich is different from implementation specific20:06
kerioyep, ##c confirms *p++ = *p is undefined20:07
freemangordonmsvc gives no warning and executes it as a[i] = a[i];i++;20:09
freemangordonok, I stand corrected20:09
freemangordon:)20:09
ShadowJKJust write what you mean instead20:10
freemangordonI already did it20:10
ShadowJK(instead of writing things that the compiler can do whatever it wants with)20:10
freemangordonbut was curious if it could be written in such a cryptic way20:10
freemangordonfor( ; i < alarm_events_cnt-1; i++)20:11
freemangordonalarms_events[i] = alarms_events[i+1];20:11
ShadowJKBtw, googling this I found one page which said that it's actually defined in java to behave not like msvc did20:12
ShadowJK(and gcc)20:12
ShadowJKlol, wikipedia says that earlier version of gcc, when encountering undefined behaviour, would try to launch nethack, rogue, or emacs20:14
kerioShadowJK: hm? i thought that was just for #pragma20:14
ShadowJKyeah20:15
*** MrPingu has joined #maemo-ssu20:15
*** toxaris has joined #maemo-ssu20:20
chem|stmerlin1991: any ideas on the gl_FUNC_ARGZ?20:32
chem|stis that needed?20:33
chem|stwas that some workaround from last time20:33
merlin1991that was the fun with libtool20:33
chem|stI broke it again20:34
chem|stwell last time I checked I had it working without any fancy hacking...20:37
chem|stI installed some python crap in between the builds20:37
DocScrutinizer05((alarms_events[i++] = alarms_events[i])) *seriously* insane, you know that20:48
*** deepy\SIGSEGV has quit IRC20:58
*** deepy\SIGSEGV has joined #maemo-ssu20:58
chem|stas long as there is this argz thing there is no modest with me no more20:59
DocScrutinizer05a coder honestly considering to ask like 100 lines of IRC questions to make sure he can use a coding shortcut abomination, really.... shoot him with owl shit21:00
DocScrutinizer05I gather freemangordon hasn't invented this obfuscated crap21:01
freemangordonit was exactly him :)21:01
chem|stpali is xterm considered stable (your last commit for portrait mode)21:05
chem|stor should I stick with the one around august?21:05
Palichem|st, patch which added new options in Font&Options dialog is stable21:10
Paliand patch which added portrait mode flag is stable too21:10
Palibut ask merlin1991 or freemangordon if we want portrait mode in xterm21:10
chem|sto.O21:11
chem|stI took "Increase version to 0.14.mh31+0cssu0" now what is the font option thing? you mean the one before with the application settings right21:13
Paliincrease version commit from Monday July 16 2012 is stable21:14
chem|stok *check* built21:14
chem|stPali: any idea what I messed up with my libtool stuff?21:15
keriohttp://www.ebay.co.uk/itm/A-piece-of-cardboard-shaped-a-bit-like-an-iPhone-5-USED-/160916451185 oh lols21:15
chem|stI know there was that issue with intltools but that one is another21:15
DocScrutinizer05freemangordon: a few basic best common practice rules about coding: Don't peephole optimize since you never can outsmart the compiler on that discipline. Never try to safe a few lines or even chars of code, when the result becomes only slightly less readable. Try to write what you planned to do, not what you think is the most terse nifty way to implement it, i.e. your code should be so clear that it's not possible to write a comment21:16
DocScrutinizer05making anything more obvious than it's in code. And final one for now: never use operators or constructs that you're not absolutely sure about what and how they do21:16
kerioDocScrutinizer05: today i discovered that a ^= b ^= a ^= b; is actually undefined :(21:16
Palikerio, why?21:18
*** arcean has joined #maemo-ssu21:18
keriosame reason, it modifies the same variable twice between sequence points21:18
DocScrutinizer05int sourceindex; int destindex; sourceindex = 1; destindex = sourceindex - 1; while (sourceindex < upperlimit) {array[sourceindex++] = array[destindex++]}21:19
DocScrutinizer05oops, other way round21:20
DocScrutinizer05{array[destindex++] = array[sourceindex++]}21:20
keriofor (int i = 1; i < limit; i++) array[i - 1] = array[i];21:20
kerioalso known as *a fucking for loop*21:20
DocScrutinizer05also known as not as clear as it could be21:21
DocScrutinizer05and you do arithmetic in a for loop which is BS (though the compiler will optimize that too, but it's not what you would do when you do it 'by hand')21:22
kerioi do what now?21:22
keriooh, i see21:22
kerioright, i forgot that arithmetic is incredibly slow, and most definetely slower than memory access21:23
keriooh wait21:23
kerioany vaguely modern cpu can just calculate i - 1 and then do nothing for a lot of time while waiting for the memory to be accessed21:23
DocScrutinizer05see point #1 of rules above21:24
kerioyour thing is peephole optimization21:24
DocScrutinizer05in fact the compiler will probably compile my code to a pretty similar assembler21:24
DocScrutinizer05nope, my thing is doing it the way a sane mind would do it manually21:25
DocScrutinizer05esp when your distance is not 1 but maybe 4, or 5521:25
kerioyou do *two* calculations in a loop21:25
DocScrutinizer05nope21:25
DocScrutinizer05increment is not a calculation21:26
DocScrutinizer05replace 1 by 55 in your and my code, and then think again21:26
kerioanyway, the right solution is obviously memmove(array + 1, array, (elements - 1) * sizeof *array)21:27
kerio:D21:27
DocScrutinizer05whatever, your code is magnitudes more sane than freemangordon's21:27
tadzikdoesn't memmove fuck up when areas overlap?21:28
keriotadzik: nope, that's memcpy21:28
tadzikah, right21:28
DocScrutinizer05and ack for memmv21:28
DocScrutinizer05though array + 1 wount work usually21:29
keriowhy not?21:29
DocScrutinizer05since you're missing out on true size of array elements21:29
kerioarrays decay to pointers to their first element when passing them or doing arithmetics with them21:29
DocScrutinizer05&array[1]21:30
kerioDocScrutinizer05: &array[1] is **effectively the same** as array+121:30
DocScrutinizer05probably not always21:30
keriohttp://c-faq.com/ptrs/explscale.html21:30
DocScrutinizer05though you *might* be right on that one21:30
DocScrutinizer05I just wonder what e.g. lint would say about it21:31
kerioK&R1 Sec. 5.3 p. 94, K&R2 Sec. 5.4 p. 103, ISO Sec. 6.3.6, H&S Sec. 7.6.2 p. 20421:31
keriohm, what's H&S?21:31
DocScrutinizer05ok ok21:31
keriooh, harbison steele21:31
DocScrutinizer05dafaq never heard of21:32
*** NIN102 has joined #maemo-ssu21:37
*** NIN101 has quit IRC21:39
*** gregoa has quit IRC21:41
MrPinguregarding xterm21:41
MrPinguSomebody noticed xterm does auto pointermode, and then it's stuck in that mode?21:41
chem|st?21:41
DocScrutinizer05still I for one prefer 'adress of array, element one' over 'arry plus one, as defined by K&R1 Sec. 5.3 p. 94, K&R2 Sec. 5.4 p. 103, ISO Sec. 6.3.6, H&S Sec. 7.6.2 p. 204'21:42
MrPinguLet me shoot video, if needed21:42
DocScrutinizer05MrPingu: I seem to recall to have seen that when 'history' aka scrollback been < one screen worth of lines21:43
DocScrutinizer05<= actually21:44
DocScrutinizer05which also applies when you go $foo mode - the one I forgot the name and being used by stuff like less, mc, top etc21:44
DocScrutinizer05if you kill -9 mc from another terminal, your screen looks almost normal until you try to scroll back21:45
DocScrutinizer05same for less, top etc21:46
DocScrutinizer05is it called block mode?21:46
DocScrutinizer05raw mode?21:46
*** gregoa has joined #maemo-ssu21:46
DocScrutinizer05nfc21:46
kerioDocScrutinizer05: splint doesn't complain about array+121:46
MrPinguWell most of the times, switching fullscreen makes the pointer icon appear again21:47
MrPinguThis time not, dafuq ;o21:47
MrPinguhmm not that21:50
kerioosso-xterm sucks21:52
MrPinguYeah :(21:52
Palinew status area applet battery is ready for testing!22:12
Palihttp://atrey.karlin.mff.cuni.cz/~pali/status-area-applet-battery_1.0-1_armel.deb22:12
Palihttps://gitorious.org/~pali/community-ssu/pali-status-area-applet-battery22:12
Palithis version is compatible with BME and also with new kernel drivers22:13
Paliusefull for usb host mode when BME is stopped and bq27x00_battery can be loaded for battery info22:13
Palimerlin1991, freemangordon, DocScrutinizer05, kerio ^^^^22:14
kerio:D22:14
kerioPali: what does it use?22:14
kerioi mean, is the switching automatic?22:14
Paliyes22:14
Palifully automatic22:14
keriocan it use bq27200 data without the modules loaded?22:15
Paliof course not22:15
kerio:(22:15
Palibut when you stop bme and load kernel drivers, you will see battery status22:15
Paliso you can use kernel driver for charging and see status in status bar22:16
kerioi'll see a *more accurate* battery status, won't i22:16
kerioi like it :322:16
Paliplus it show percentage, current and design capacity and time until battery will discharge22:17
Pali(or until fully charge)22:17
kerioPali: now add a button for host mode! :D22:18
Palikerio, host mode button?22:18
Paliit is in usbmode package22:18
kerioyeah but that's a *new* button22:19
Palithere is no place for button in battery status plugin22:19
Pali:D22:19
kerioof course there is, *the whole thing*22:19
kerio:322:19
kerioPali: how do i make the battery indicator show up in the status area without rebooting? :s22:20
kerioit's not hildon-home22:20
MrPingustatusmenu, i guess22:20
kerioMrPingu: ty22:20
*** arcean has quit IRC22:20
Pali$ killall -9 hildon-status-menu; sleep 1; killall -9 hildon-status-menu; sleep 1; killall hildon-status-menu22:21
Palikerio ^^^22:21
kerioPali: hm, can the data from bme be trusted?22:21
MrPinguit's working :)22:21
kerioyeah, it works here too22:21
Palisome plugins sucks in my n900, so need to kill it twice with -9 and once normal22:21
MrPingubme is some kind of mystery :P22:22
*** gregoa has quit IRC22:22
Palikerio, I design capacity from bme is same as from rx51_battery driver (+/- small epsilon)22:22
kerioPali: what are the modules to load again?22:22
*** gregoa has joined #maemo-ssu22:22
Palirx51_battery for design capacity22:22
Pali(this can be loaded when bme is running too)22:23
kerioit's already loaded22:23
keriobut it's not as accurate as the bq27200 data, right?22:23
Paliand bq2415x_charger & bq27x00_battery when BME is stopped22:23
Palibq2415x_charger is needed only if you want to charge battery22:24
Palistatus menu plugin do not need it22:24
kerioconfirmed working :D22:24
Palistatus menu plugin needs bq27x00_battery (+ rx51_battery for design capacity) or BME22:24
Palikerio, ok :-)22:24
kerioPali: hold on, it doesn't take the capacity from bq27200? :(22:25
keriowhen available, i mean22:25
Palikerio, it take current mAh from bq2700 if possible22:25
MrPinguNever bothered with those bqxx battery things, but BME seems to report fine22:25
MrPinguPali, does it take the default graphics?22:26
kerioPali: my bq27200-0/charge_full is 126163822:26
kerioapplet reports 126722:26
kerioit doesn't seem right :s22:26
Palikerio, then it take design mAh from rx51_battery (if not possible then take last full mAh from bq27200) and calculate percentage22:26
Palikerio, it report design capacity if possible from rx51_battery22:26
kerio...why? :S22:27
Paliif not then it take last full22:27
Paliwhy not?22:28
keriobecause bq27200 is accurate, rx51_battery is not22:28
Palibq27200 do not know design capacity...22:28
Paliand rx51_battery reporting now design capacity22:29
keriodesign capacity doesn't mattery, what matters is the biggest charge it can hold22:29
PaliMrPingu, yes it have default graphics22:29
Palikerio, but there is not way how to get biggest charge22:29
kerio/sys/class/power_supply/bq27200-0/charge_full22:30
Palithis is capacity after *last* charge22:30
Palinot biggest22:30
kerioafter the last calibration, you mean22:30
*** MrPingu has quit IRC22:31
keriowhich is what we should use, if it's available22:31
kerioDocScrutinizer05: confirm/deny?22:31
Palikerio, if you really want to use that value rmmod rx51_battery22:32
keriothat's what i'm doing :)22:32
Palistatus menu plugin will switch to use it :-)22:32
keriobtw, what's the status of the bme userland replacement?22:32
Palimaybe it is time to try it :-)22:33
Palihald-addon-bme worked when I tried it last time22:33
Paliand libbmeipc worked when freemangordon tested it22:34
keriohm22:34
PaliI only added temperature support from rx51_battery to libbmeipc but not tested22:34
*** dhbiker has joined #maemo-ssu22:35
Palikerio, do you want to try hald-addon-bme?22:37
kerioyeah, but let me make a full backup first22:37
Paliok22:38
*** _ade_ has joined #maemo-ssu22:44
*** MrPingu has joined #maemo-ssu22:45
_ade_Pali: perhaps you should make your status-area-applet-battery conflict with Advanced Power? I had a battery icon twice in the status menu and a double battery status in the menu after installing22:46
Pali_ade_, my version replace stock package22:47
Paliit is same as if you reinstall stock nokia package22:47
Palithen it is problem with advanced power...22:47
_ade_pali: I guess you are right if you put it that way.22:47
kerio_ade_: reinstalling APM will probably fix it22:48
kerioalso try rebooting22:48
kerioor killing hildon-status-menu22:48
_ade_kerio: I removed advanced power before rebooting22:49
_ade_will try to reinstall APM...22:49
kerioPali: ok, how do i install your bme replacement?22:50
Palikerio, full update of packages I did not tested22:51
PaliI only tested hald-addon-bme and now I commited more patches to it22:51
keriowell, the bme package should be fairly solid, right?22:51
kerioit's just a replacement of initscripts22:51
PaliI will push all changes to git22:52
_ade_pali: reinstalling APM worked. It does not revert any of the changes you made?22:56
Pali_ade_ I do not know what APM is doing...22:57
*** MrPingu has quit IRC22:57
_ade_Me neither ;-)22:57
Palimy status area plugin is in same package name as Nokia22:57
Paliand has same file name22:57
Palionly version was increased :-)22:57
Palikerio, now all changes are in git on: https://gitorious.org/rx51-bme-replacement/22:57
Palikerio, if you want to test hald-addon-bme: backup old hald-addon-bme binary and update it. then run: sudo stop hal; sudo /etc/init.d/hal stop; sudo /etc/init.d/hal start22:59
keriodoes the new hald-addon-bme work with bme too?23:00
Palino23:00
kerioor, at least, does it stay quiet if bme is running?23:00
Paliit work only with kernel drivers23:00
Paliyes, it can be started, but it report nothing23:00
_ade_Pali: then I think I will stick you version solely (without APM) and see if I encounter issues23:01
Palikerio, and it can wait for hotplug kernel drivers23:01
Paliand it write very long log output to: /tmp/hald-addon-bme.log23:01
Pali_ade_, ok23:02
Palikerio, and hald-addon-bme binary is in /usr/lib/hal/23:05
*** toxaris has quit IRC23:11
*** MrPingu has joined #maemo-ssu23:18
*** gregoa has quit IRC23:18
*** gregoa has joined #maemo-ssu23:20
*** MrPingu has quit IRC23:23
*** MrPingu has joined #maemo-ssu23:24
Palikerio, _ade_, if you testing new battery status area applet, tell me if all notifications (battery low, battery empty, battery full, charging, discharging) working23:29
Palialso if sound notifications too23:29
_ade_Pali: okay23:29
Paliand if no sound play when silent profile (or disabled system sounds)23:29
*** lizardo has joined #maemo-ssu23:32
*** arcean has joined #maemo-ssu23:32
Palikerio, hald-addon-bme working23:33
kerioare you asking me?23:33
*** NIN102 has quit IRC23:34
kerioPali: also, what deals with the led and the shutdown on EDVF?23:35
kerioPali: also i don't have a toolchain to compile hal-addon-bme :)23:36
Palihald-addon-bme managing MCE led patterns and also send DSME signal for shutdown23:36
keriodoes hald-addon-bme need rx51-battery?23:36
Paliyes, for design capacity23:37
Palibut working without it23:37
Pali(it report design capacity = 0)23:37
Palikerio, here is elf binary: http://atrey.karlin.mff.cuni.cz/~pali/hald-addon-bme23:37
_ade_Pali: should the battery status now still work once bme is stopped? Or does an alternative driver has to be loaded then?23:38
kerio_ade_: you need the modules23:38
Pali_ade_, if bme is stopped you need to load bq27x00_battery module23:38
keriobq2415x-charger and bq27x00-battery23:38
keriooh, maybe you don't need -charger23:38
Paliand bq27x00_battery module must be unloaded before bme is started23:38
_ade_okay, I thought so, but just checking.23:39
kerioPali: hm, stopping hal caused my n900 to shutdown23:40
Palikerio, how did you stopped hal?23:41
keriostop hal23:41
Palihm...23:41
kerioanyway, i'm going to try charging23:42
kerioPali: hahaha wtf did you do23:43
kerioi get the charging message when the charger isn't connected23:43
Palifrom HAL?23:44
keriook, now it's working correctly23:44
keriofor a moment it was like inverted23:44
keriook, i'm getting the correct LED pattern when the screen is locked23:45
kerioPali: it's horrendously slow at figuring that the charger was plugged in/removed23:46
Palikerio, I see that too23:46
Palitrying to debug it23:46
kerioPali: it's just the status menu i think23:46
Palino23:46
Paliproblem is in hal23:46
keriothe led pattern is instant-ish23:46
Paliah... there is another problem: status menu plugin will not read values from bme if bme is not running23:47
Palior new version of bme package is not installed23:47
Paliso status menu plugin getting charger value from bq27x00_battery23:47
Paliand this can take a time23:47
keriohm23:47
keriois there no way to make that faster'23:48
kerio?23:48
Paliso I need to add 1) bq2415x_charger support to status menu or 2) use values from HAL bme if bme is stopped or 3) update bme package :-)23:48
Palikerio, I think that option 3) will be enought23:49
Palikerio, ok problem is not in hal, only in status menu plugin23:52
keriowhat does "update bme package" mean?23:53
keriothat is, what will it change?23:53
Paliupdate bme-rx-51 deb package23:53
kerioyeah, but what will change?23:54
keriowhy would it make the battery applet update more quickly?23:54
Paliif bme-rx-51 deb package version 1.0 (or high) is installed, then status menu plugin thinks that BME is running23:54
keriooic23:54
Paliif status menu plugin thinks that bme is not running it will not read values from hald-addon-bme23:54
Pali(because that values can be out-of-date)23:55
keriohold on, is the check actually on the version on the package? :S23:55
keriojust... no23:55
Paliyes, on the package version23:55
kerio:S23:56
Palithis is good way how to check if new version of binary in debian package is installed23:56
keriohold on, new version of the binary? :o23:56
kerioyour bme actually ships a bme binary?23:56
Palino bme binary23:57
Palionly bme-rx-51 deb metapackage23:57
keriocan't you ship an empty /etc/nobme file instead?23:57
kerioif anything, because checking a package version is quite expensive23:58
Paliwhy? it is check once at startup23:58
kerio...oh, so that's why the status menu hangs for like 5 seconds before loading >:C23:58
Palinow it show you message when battery is really started charging :-)23:59
keriodepending on dpkg is an unnecessary complication23:59

Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!