IRC log of #maemo-ssu for Friday, 2013-11-08

*** Pali has quit IRC00:28
RaimuYeah, it was 10 hours to go but I guess the hour changed at a funny point.00:37
*** dhbiker has joined #maemo-ssu00:44
sixwheeledbeastweird00:45
*** hbib has quit IRC00:48
*** KenYoung has joined #maemo-ssu00:54
*** LauRoman has joined #maemo-ssu01:13
DocScrutinizer05hihi01:16
DocScrutinizer05this happens when coder doesn't obey one of golden rules: always call timing/metering functions only once, in an atomic way, then parse result and store to local vars01:18
DocScrutinizer05if you however do crimes like   printf("time: %2d:%2d", ftime() / 3600, ftime() - (ftime() / 3600))  [or similar] then you get results like the one sixwheeledbeast reported above01:24
KenYoungThat's *really* bad code.01:25
DocScrutinizer05s@ftime() - (ftime() / 3600)@(ftime % 60)/6001:25
*** okias has quit IRC01:25
DocScrutinizer05KenYoung: it's incorrect, but even when it was correct it still would be pretty BAD01:26
DocScrutinizer05ftime() returns seconds,milliseconds since epoch, so the code above is garbage, but I just tried do illustrate something01:27
*** M4rtinK has joined #maemo-ssu01:29
DocScrutinizer05actually it *returns* always 001:29
*** okias has joined #maemo-ssu01:29
DocScrutinizer05the point however is: you NEVER call a readout function of a dynamic value multiple times. You always call it *once* and store the returned value to a local var01:32
*** okias has joined #maemo-ssu01:32
DocScrutinizer05obviously the coder of the app that created that notifier sixwheeledbeast seen didn't know that01:32
DocScrutinizer05it could even get worse when the app hits a microsecond earlier then it may display 10:60 instead 10:00, a bit later and it displays 9:00 instead 10:0001:35
DocScrutinizer05on driver level when you can't do atomic reads (e.g. on cmos clock) you do two reads of all registers and compare. If they differ you read a third time, and so on01:36
*** kolp has quit IRC01:43
*** obiwlan has quit IRC01:48
*** M4rtinK has quit IRC01:49
*** obiwlan has joined #maemo-ssu01:50
*** amizraa has joined #maemo-ssu01:59
*** Martix has joined #maemo-ssu02:34
*** okias has quit IRC02:45
*** Martix has quit IRC02:48
*** nox- has quit IRC02:57
*** DrCode has quit IRC03:07
*** LauRoman has quit IRC03:13
*** DrCode has joined #maemo-ssu03:15
*** dos1 has quit IRC04:31
*** amiconn_ has joined #maemo-ssu05:52
*** amiconn has quit IRC05:52
*** amiconn_ is now known as amiconn05:52
*** M13 has joined #maemo-ssu06:33
*** trx has quit IRC07:05
*** KenYoung has quit IRC07:54
*** oldtopman has quit IRC08:27
*** Pali has joined #maemo-ssu08:35
*** Pali has quit IRC09:18
*** LauRoman has joined #maemo-ssu09:24
sixwheeledbeastDocScrutinizer05:  Raimu spotted it, I thought it would be impossible to see. However as you point out bad coding somewhere which I guess many will be unsurprised10:00
*** M13 has quit IRC10:08
Raimuyup10:15
DocScrutinizer05my fault10:23
RaimuThat code should be revisited considering it's in a CSSU replacement block, but it's a relatively "huh." level thing, thankfully caught.10:24
*** LauRoman has quit IRC10:37
*** Martix has joined #maemo-ssu10:46
*** trx has joined #maemo-ssu10:59
*** kolp has joined #maemo-ssu11:00
*** chem|st_ is now known as chem|st11:03
*** Martix has quit IRC11:15
DocScrutinizer05in which component is this?11:17
*** Martix has joined #maemo-ssu11:18
*** tom_ has quit IRC11:20
freemangordonmerlin1991: going to push libglib in -devel, should I add +0cssu0 suffix to the version?11:23
*** Martix has quit IRC11:24
*** RST38h has quit IRC11:50
*** RST38h has joined #maemo-ssu11:50
*** M4rtinK has joined #maemo-ssu11:54
*** Martix has joined #maemo-ssu12:08
*** Pali has joined #maemo-ssu12:15
*** arcean has joined #maemo-ssu12:19
*** tom__ has joined #maemo-ssu12:20
*** trx has quit IRC12:35
*** lizardo has joined #maemo-ssu12:35
*** delphi has joined #maemo-ssu12:35
*** Martix_ has joined #maemo-ssu12:37
*** Martix has quit IRC12:40
*** Martix_ has quit IRC12:45
DocScrutinizer05DANG, seems MohammadAG is absolute top in Itt-thanks (karma) even before qgil, texrat, stskeeps, freemangordon12:57
DocScrutinizer05:-D12:57
DocScrutinizer05but wait, I'm coming12:58
DocScrutinizer05http://maemo.org/profile/list/category/itt_thanks/12:58
DocScrutinizer05i'm in good neighbourhood of Kash and Andrew, but we're hunting E*_ :-O13:00
DocScrutinizer05Pali: I probably should style-prrofread your tmo posts ;-) otherwise E*_ will leave you behind13:01
RaimuDocScrutinizer05: Pretty sure the 9:60 bug is in clock-ui13:23
*** Martix_ has joined #maemo-ssu13:28
*** Martix_ has quit IRC13:33
*** Martix_ has joined #maemo-ssu13:35
sixwheeledbeastRaimu: clock-ui or open source qt replacement clock-ui?13:39
*** delphi is now known as trx13:39
* DocScrutinizer05 senses a fundamental problem in maemo: missing peer review on source level, of *all* code before it hits the user13:40
DocScrutinizer05pretty unusual and a nogo for industrial grade software - like original maemo fremantle (at least I'd hope for that in original fremantle as well, we definitely had it in modem firmware stack development for ALL code)13:42
*** Martix_ has quit IRC13:43
*** Martix_ has joined #maemo-ssu13:43
DocScrutinizer05for kernel (and mainline system) you got that peer review on public ML where you publish your patches to get them merged13:50
DocScrutinizer05for maemo core programs we need sth similar13:50
DocScrutinizer05no matter which flavor/branch of maemo system, be it CSSU or even core apps shipped via extras13:51
DocScrutinizer05testing QA is *way* too weak13:52
*** l4m3rx has quit IRC13:57
DocScrutinizer05aux info: we planned 1..2h for the peer review meeting (incl notes to take, to prepare proper protocol later on) and the reviewer is supposed to use at least 4..8h to review the code before discussing it with original author in review meeting14:01
DocScrutinizer05depending on result of the review meeting, a second one might be needed to review the changes that resulted from first one14:02
Raimusixwheeledbeast: The replacement.14:03
sixwheeledbeastit's bad enough getting enough testers to QA extras-testing packages.14:03
DocScrutinizer05and it's strictly forbidden to get "help" from original author during the review task, since otherwise your view on stuff gets infested with a possibly false explanation on how stuff works, given by the original author14:04
sixwheeledbeastRaimu: please report to ade then so it can be fixed. http://talk.maemo.org/showthread.php?t=81582 http://wiki.maemo.org/Clock_replacement14:05
DocScrutinizer05sixwheeledbeast: that's not a sufficient excuse to do sloppy QA14:05
sixwheeledbeastdidn't say it was.14:05
DocScrutinizer05sixwheeledbeast: that's at best an excuse why QA takes ages until you find somebody willing and able and with the right schedule14:06
sixwheeledbeast^^ was just going to say the same14:06
DocScrutinizer05I'd like to suggest to create a new rule: no package/patch goes to CSSU-S without proper *documented* sourcecode peer review14:07
DocScrutinizer05and a rule/guideline from industry: you may have more than just one final peer review. Actually you should discuss stuff early and frequently with your peers14:09
DocScrutinizer05often even a developer team doing an almost continuous source review is more productive than two developers working solitary on their respective tasks14:10
sixwheeledbeastnew CSSU-S is needed soon IMO. There are fixes to important things needed in Stable. A-GPS fix, hildon-desktop rotation,  etc..14:11
DocScrutinizer05that however is pretty much depending on the individuals and how they "work together"14:11
sixwheeledbeast:nod:14:11
DocScrutinizer05we should encourage teambuilding like this14:12
DocScrutinizer05making peer review mandatory for CSSU-S is a first good step to encourage teambuilding14:12
DocScrutinizer05yep, GPS fix is pretty important14:13
DocScrutinizer05what's the rotation issue?14:14
*** LauRoman has joined #maemo-ssu14:26
*** dos1 has joined #maemo-ssu14:31
*** dos1 has quit IRC14:41
*** trx has quit IRC14:42
*** dos1 has joined #maemo-ssu14:42
*** trx has joined #maemo-ssu14:46
*** trx has joined #maemo-ssu14:46
*** sono has quit IRC14:47
*** sono has joined #maemo-ssu14:47
*** Martix_ has quit IRC14:49
*** okias has joined #maemo-ssu14:51
*** KenYoung has joined #maemo-ssu14:52
*** freemangordon_ has joined #maemo-ssu15:15
*** freemangordon has quit IRC15:15
*** KenYoung has quit IRC15:31
*** Martix_ has joined #maemo-ssu15:42
*** NIN101 has joined #maemo-ssu15:48
*** Martix has joined #maemo-ssu15:52
*** l4m3rx has joined #maemo-ssu15:52
*** Martix_ has quit IRC15:55
*** Martix has quit IRC15:57
*** Martix has joined #maemo-ssu16:00
*** lizardo has quit IRC16:31
*** dhbiker has quit IRC16:36
*** lizardo has joined #maemo-ssu16:36
freemangordon_merlin1991: ping16:41
*** freemangordon_ is now known as freemangordon16:42
merlin1991pong16:42
freemangordonmerlin1991: stock glib puts .so in /usr/lib, 2-24 in /lib16:42
freemangordonis that a problem?16:42
freemangordonactually it puts libglib-2.0.so.0.2400.1 in /lib, all others in /usr/lib16:43
freemangordonlike libgio, etc16:43
merlin1991hm do you think any binary has libglib linked with an rpath?16:43
freemangordonit would be a madness16:43
merlin1991yeah :D16:44
*** NIN101 has quit IRC16:44
merlin1991 /lib is inside the serarchpath for libs so this *should* not affect stuff16:44
*** NIN101 has joined #maemo-ssu16:44
freemangordonsure16:44
freemangordonbut I wanted a confirmation :)16:44
freemangordonactually afaik /lib should be before /usr/lib16:45
freemangordonand it is mounted (or should be) earlier in the boot process16:45
merlin1991I guess that is the reason glib is there16:45
merlin1991since a lot of stuff nowadays depends at least on the memory stuff glib provides (lists and such)16:46
freemangordonyep, but on stock it is in /usr/lib. however, I'll leave it that way, will fix it if there is any problem16:46
freemangordonthough I doubt there will be16:46
*** arcean_ has joined #maemo-ssu17:04
*** arcean has quit IRC17:07
*** Martix has quit IRC17:23
*** Martix has joined #maemo-ssu17:37
sixwheeledbeastRaimu: Thanks for bug report. I have added some links to IRC logs18:31
*** dos1 has quit IRC18:34
*** dos1 has joined #maemo-ssu18:39
sixwheeledbeastDocScrutinizer05: This bug fixed in Jan 2013 on CSSU-T >>> http://talk.maemo.org/showpost.php?p=1315898&postcount=1117 http://talk.maemo.org/showpost.php?p=1315929&postcount=112418:43
sixwheeledbeastI am sure there are others.18:43
*** dos1 has quit IRC18:47
*** Martix has quit IRC18:53
arcean_omg18:56
arcean_sixwheeledbeast: are you it's notfixed in stable?18:56
*** nokiabot has joined #maemo-ssu19:22
nokiabotwhere am i ??19:23
*** nokiabot has quit IRC19:23
DocScrutinizer05LOL19:26
sixwheeledbeastarcean_: AFAIK hildon-desktop on CSSU-S is later than your patch to fix above.19:27
sixwheeledbeastI still have this happen every now and then on CSSU-S.19:29
*** dos1 has joined #maemo-ssu19:30
*** dos1 has quit IRC19:31
*** dos1 has joined #maemo-ssu19:31
*** dos1 has quit IRC19:43
*** dos1 has joined #maemo-ssu19:48
*** oldtopman has joined #maemo-ssu20:24
*** Martix has joined #maemo-ssu20:58
*** Woody14619a has joined #maemo-ssu21:41
*** Woody14619 has quit IRC21:44
*** Woody14619a is now known as Woody1461921:47
*** Woody14619a has joined #maemo-ssu22:33
*** Woody14619 has quit IRC22:35
*** Martix has quit IRC22:56
*** lizardo has quit IRC22:57
*** Martix has joined #maemo-ssu22:59
*** arcean_ has quit IRC23:03
*** obiwlan has quit IRC23:08
*** obiwlan has joined #maemo-ssu23:10
*** _rd has joined #maemo-ssu23:27
*** nox- has joined #maemo-ssu23:56

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