IRC log of #maemo-devel for Thursday, 2010-02-18

« Wednesday, 2010-02-17 Index Friday, 2010-02-19 »
ifreqhowto promote from dvel to testing?00:11
lcukgo to the packages page for your app, whilst logged in00:11
lcukand theres a button/link/chicky thing afaik00:11
ifreqhmm ok sec00:11
ifreqhmm does the testing repo need bugtracker link also?00:15
woglindeyeah00:15
lcukifreq, not sure00:15
woglindewere should the testers send bugreports?00:15
ifrequhm ok00:15
ifreqneed to fix it then00:16
ifreqtoo bad bugs.maemo.org userlogin is not share with garage :P00:17
ifreqagain new acc to make :)00:17
ifreqhmm mayb its easier for me to use emailaddr on bugtracker field 1st00:29
*** BabelO has quit IRC01:04
*** lardman has joined #maemo-devel01:05
*** t-tan has quit IRC01:32
*** VDVsx has quit IRC01:43
*** SpeedEvil1 has joined #maemo-devel01:46
*** SpeedEvil has quit IRC01:46
*** lbt has quit IRC01:51
*** SpeedEvil1 is now known as SpeedEvil01:55
*** lardman has quit IRC01:58
*** rsalveti has quit IRC02:35
*** rsalveti has joined #maemo-devel02:36
*** woglinde has quit IRC02:36
*** SpeedEvil has quit IRC02:49
*** SpeedEvil has joined #maemo-devel02:49
*** god206 is now known as swc|66603:09
*** edisson has quit IRC03:24
*** nasrudin has joined #maemo-devel03:53
nasrudingood night guys. Just a quick question, I was trying to grab a signal via dbus and alter some of it's data but it looks like I can just listen to the bus and not alter stuff if I am not the receiver. Any ideas on how to accomplish that?03:55
nasrudinwell it boils down to this, I want to grab all the calls made and alter the number those calls are using. Maybe there's a smarter way than via d-bus...04:04
*** memble- has joined #maemo-devel04:55
*** memble- is now known as mboh04:58
*** swc|666 has quit IRC05:09
*** nasrudin has quit IRC05:10
*** Acedip has joined #maemo-devel05:35
DocScrutinizerat least dbus lacks the concepts to do what you sketched, it seems to me05:50
SpeedEvilchroot the source or destination.05:51
SpeedEvilRun a dbus server in the chroot05:51
DocScrutinizeryou'd need a 'MITM' and separate sender and receiver on different busses (or users)05:52
SpeedEvilyeah05:53
*** jebba900 has joined #maemo-devel06:02
*** jebba900 has quit IRC06:22
*** slonopotamus has quit IRC07:01
*** slonopotamus has joined #maemo-devel07:04
*** DocScrutinizer has quit IRC07:30
*** DocScrutinizer has joined #maemo-devel07:30
*** ptl has quit IRC08:33
*** ptl has joined #maemo-devel08:35
*** ptl has quit IRC08:35
*** ptl has joined #maemo-devel08:35
*** tekojo has joined #maemo-devel08:48
*** madsy has quit IRC08:56
*** bricks has joined #maemo-devel09:01
slonopotamushmm09:03
tekojoooh09:08
* slonopotamus hack, hack, hack...09:17
*** fragment_ is now known as fragment09:17
ptlhssudsuds09:26
ptlsorry09:26
ptlI was yawning.09:26
*** tbf has joined #maemo-devel09:27
*** lbt has joined #maemo-devel09:33
*** BabelO has joined #maemo-devel09:33
*** cpm has joined #maemo-devel09:41
cpmHi, is there a tun device under Maemo4? When my code wants to open /dev/net/tun the kernel seems to want to load the kernel module but fails. Is there generally tun support?09:42
*** tbf has quit IRC09:49
*** ppenz has joined #maemo-devel09:59
*** Modeless has joined #maemo-devel10:06
slonopotamusis there any reason to deny directory listing http://repository.maemo.org/dists/fremantle/sdk/free/ ?10:12
*** tbf has joined #maemo-devel10:14
*** Modeless has quit IRC10:20
*** amigadave has joined #maemo-devel10:35
*** guido_g has joined #maemo-devel10:43
*** guido_g has left #maemo-devel10:43
*** ph1l has quit IRC11:33
*** sleipnir has joined #maemo-devel12:04
*** tekojo has quit IRC12:12
*** tekojo has joined #maemo-devel12:13
*** lardman has joined #maemo-devel12:20
*** t-tan has joined #maemo-devel12:20
*** ph1l has joined #maemo-devel12:21
*** yerga has joined #maemo-devel12:28
*** dazo_afk is now known as dazo12:29
lardmanhmm, I love "your app sux" type comments on TMO, makes me wonder why I even bother with users12:38
fralslardman: know the feeling12:39
*** mboh has quit IRC12:43
*** ppenz has quit IRC12:44
*** ppenz has joined #maemo-devel12:45
*** mboh has joined #maemo-devel13:07
*** amigadave has quit IRC13:17
*** edisson has joined #maemo-devel13:31
*** VDVsx has joined #maemo-devel13:35
*** polac has quit IRC13:50
*** edisson has quit IRC13:51
*** polac has joined #maemo-devel13:52
dazolardman:  at least it is a feedback ... not much useful feedback ... but it's better with some feedback than none at all :)13:58
lcuklardman, frals, a good intermediary dev/tester liaison team would be handy in many regards14:00
fralslcuk: indeed14:00
lcuksame team as the testers14:00
lcukerrr bug squad14:00
lcukpeople who know the devs and the users :)14:01
lcukits important for you guys to know when theres problems but not to be bugged every 2 seconds by them14:01
lardmanit's not the bugging, just the offhand abuse14:01
lardmannot very constructive, it's crap, do it better14:02
lardmanand I think, I'm not writing it for you, piss off :)14:02
lcuklol14:02
*** tekojo has quit IRC14:06
*** amigadave has joined #maemo-devel14:06
fralsthe problem i have is users posting "not working" and not bothering to read the previous posts where im constantly asking for logfiles so i can actually track down the cause of it14:07
fralsdoing my head in14:07
lcuksplash14:10
lcukpeople will not read forums every time14:10
lcukshow a banner14:10
lcukdialog14:10
lcuksomething14:10
lcukwhilst its in testing14:10
lcuknoting the current expected flaws14:10
lcukput it in the users face14:11
lcukon the device14:11
fralslcuk: tried, it failed14:11
fralsi put the yellow info banner you get after installing something via ham14:11
fralsthat you need to actually click to get rid off14:11
fralsand ppl still didnt read it14:11
lcuknahh not quite what i mean14:12
lcukbut closer than nothing14:12
lardmanWhat I dislike is the attitude, they expect things to work and for things to be fixed asap14:16
lardmanfine if they pay me money, but not for free sw.....14:16
lcukyeah14:16
lcukbtw, did you solve your grid vs ownmade problem14:16
lardmanI'll stick with the table, just need to find some time to implement it14:19
lcuklardman, good idea for now, using the existing component will mean less surprises later (hopefully!)14:24
*** rsalveti has quit IRC15:02
*** achipa has joined #maemo-devel15:04
*** _Lucretia__ has joined #maemo-devel15:11
*** _Lucretia_ has quit IRC15:12
*** edisson has joined #maemo-devel15:18
cpmHi, is there a tun device under Maemo4? When my code wants to open /dev/net/tun the kernel seems to want to load the kernel module but fails. Is there generally tun support?15:22
fluxI used openvpn in n810, so there must be15:25
fluxperhaps separately packaged?15:25
flux(keeps working great on n900 as well)15:25
cpmflux: thanks for the clarificationn15:33
*** dazo is now known as dazo_afk15:34
*** aboyer_ is now known as aboyer15:45
*** t-tan has quit IRC15:52
*** rsalveti has joined #maemo-devel15:59
*** lcuk2 has joined #maemo-devel16:23
*** w00t has quit IRC16:26
*** lcuk has quit IRC16:26
*** w00t_ has joined #maemo-devel16:26
*** dazo_afk is now known as dazo16:29
*** vlad has joined #maemo-devel16:43
*** Docscrutemp has joined #maemo-devel16:47
*** Docscrutemp is now known as DocScrutinizer5116:47
*** t-tan has joined #maemo-devel16:48
*** t-tan has quit IRC16:48
*** kinta has joined #maemo-devel16:59
*** vlad has quit IRC17:00
*** _Lucretia__ has quit IRC17:04
*** Modeless has joined #maemo-devel17:08
*** cpm has quit IRC17:10
*** _Lucretia__ has joined #maemo-devel17:16
*** Modeless has quit IRC17:18
*** kodomo has joined #maemo-devel17:40
*** kodomo_ has quit IRC17:44
*** kinta has quit IRC17:50
*** joejoe has joined #maemo-devel18:13
*** joejoe has quit IRC18:18
*** v13 has joined #maemo-devel18:28
* v13 needs pyqt help18:30
*** amigadave has quit IRC18:32
v13I've a .py that uses pyqt4 and segfaults. I've found a one-line change that fixes the segfault, but it doesn't make sense. Any ideas ?18:34
HukkaWrite a bug report18:34
v13Perhaps it's my fault18:35
HukkaIF it really doesn't make sense :)18:35
HukkaI don't know, I haven't seen the change18:35
v13I can give you the .py if you want... just to be able to test it.18:35
v13in short, if i keep a reference to a QDialog it doesn't crash.18:35
HukkaOr at least paste the code18:35
v13it crashes otherways.18:35
v13It's 300 lines18:35
v13one mom18:36
HukkaBarely the size of small gif, then18:36
v13not good for paste :_)18:36
HukkaWell I didn't mean paste here :o18:36
HukkaBut in pastebin etc18:36
v13oh.. goody.. i made it seg-fault in my pc too... that's new :)18:37
v13can you access: http://v.it.teithe.gr/~v13/maemo/ ?18:37
HukkaNope18:38
HukkaShould I?18:38
v13most probably not.18:38
v13here it is: http://aetos.it.teithe.gr/~v13/maemo/18:38
v13the win.py18:38
HukkaAnd the relevant lines?18:39
v13if you remove the comment from line 210 it doesn't crash18:39
v13it runs from PC with PyQt too18:39
v13press "channels" select an entry, repeat as needed18:40
v13(that's the result of 3 full-days of debugging)18:41
v13did it crash ?18:41
HukkaUgh, I just realized that you are using mixed whitespace18:42
HukkaMakes reading that code on my setup nigh impossible18:43
Hukkav13: Never mix whitespace in Python!18:43
v13?18:43
v13well.. 8-chars tab is a bit long18:43
v13and I use vi which handles it perfectly :)18:43
HukkaSo why didn't you set your tabs to intend less?18:44
v13it messes everything18:44
v13you see different things on different editors/places18:44
v13well.. you can always replace tabs with 8-spaces18:45
HukkaI did18:45
v13 :%s/^V^T/        /g18:45
HukkaJust saying that you have a bad habit18:45
v13oh :)18:45
v13no i don't :P18:45
v13what editor do you use ?18:45
Hukkavim18:45
v13eh?18:45
v13and you have problems with vim ?18:45
v13i use vim too18:45
HukkaNo, I have problems with your code. Listen, not everybody has tab set to present eight chars. That's just how you set your's18:45
v13at the end there is a modeline18:45
v13oh18:46
v13you can always copy-paste the modeline at the end18:46
HukkaSo use either pure tabs or pure spaces18:46
v13or enable modelines (insecure)18:46
v13just do this: set ts=8 sts=4 sw=4 noet formatoptions=r ai nocindent18:46
v13you'll have the same environment as me after that :)18:47
HukkaAnd you think that's the answer? Everybody should use your env?18:47
HukkaInstead of you following PEPs and doing pure whitespace18:47
v13eh?18:47
v13which PEP says that?18:47
Hukka*sigh*18:47
v13818:47
HukkaOf course 818:47
* v13 not a py-guru18:48
HukkaYou knew enough to guess the pep18:48
HukkaNow read it.18:48
HukkaOff you go, now. And while you're gone, I'll see about your code18:48
v13however, to my defense, I've a proper modeline included in the file that can be used by any vim instance if modelines are enabled18:48
v13i didn't guess the pepe18:48
v13i've read some peps18:48
HukkaNot everyone uses vim, so the modeline argument is moot. Now go, or I will get bored with this and never get to the part where I read the code!18:49
v13eh?18:49
v13i'm reading it18:49
v13the fun thing is that I'm somehow pop 8 compliant :)18:51
v13well.. it seems that i'm compliant with the parts that are common with c programming18:52
v13did it crash ?18:53
*** wirelessdreamer has quit IRC18:54
HukkaDidn't bother installing pyqt yet. I'm looking what you are doing first18:54
v13i can save you some time on that18:54
HukkaSince if the reference isn't left anywhere, it's quite obvious that it crashes18:54
v13not exactly18:55
v13you can add a parent to the dialog and still get crashes18:55
v13I can make this crash with a number of ways, including the way it doesn't crash now.18:55
v13IOW: I have a version that crashes even without that comment.18:56
v13and ... fwiw ... it shouldn't segfault... right ?18:56
HukkaWell, the bindings area bit... quirky18:56
HukkaAFAICS you are not saving that dialog anywhere18:57
HukkaYou need a reference to it somewhere in your Python code18:57
HukkaNot saying that's a sensible from the bindings, but that's how it is18:57
v13self.slotButtonWhatDialog18:57
v13it's saved there18:57
v13and it is never used when that object is destroyed18:57
HukkaIn the code I have, you set that to None in line 21118:57
v13FCK18:58
v13anyway.. that's not the problem18:58
v13one mom18:58
v13it also crashes if you comment this out18:58
HukkaThis, you mean that line 211?18:59
v13yes18:59
v13it just needs (at-least) one more iteration18:59
v13(click-select)18:59
HukkaWell, I haven't exactly followed the code path, but I see that you set that to none in many other places too19:00
HukkaJust saying, if you make sure there's a reference to it, it should work19:00
HukkaThe segfault is very most likely from the fact that something gets gced before it should be19:00
v13no i don't19:01
v13there no other such place except from the constructor19:01
HukkaIt's not enough to give it through the bindings19:01
v13 /WhatDialog.*None will show you that19:01
HukkaIf the python side cleans it, problems ensue19:01
v13look.. The object "ui" which is an instance of Ui_Time or Ui_Channels creates the dialog and uses it19:01
HukkaAnd that's what the commented line doe19:02
Hukka+s19:02
v13if ui is deleted then the dialog is never accessed any more... another one is created19:02
v13please read what I'm writting19:02
v13I keep a reference in self.buttonWhatDialog19:02
v13and it is only set to None in the constructor19:02
HukkaI haven't read the whole code, I'm just speculating what might be the problem19:02
v13so I *keep* a reference to the dialog all the time19:02
v13As I said, I'm debugging that for 3 days19:03
v13That's why i added the global "old" to keep a reference19:03
v13but no-good19:03
Hukkano-good? Didn't you say that doing that removes the problem?19:03
v13if you leave the old.append() uncommented *and* comment the next line (211) it also crashes19:04
v13yes.. that's what it seams19:04
v13seems19:04
v13but commenting the next line it re-introduces it19:04
*** bricks has left #maemo-devel19:04
v13one mom to test that again19:04
HukkaOk, so what steps I need to do to crash it?19:06
v13Here are two ways:19:06
v13(a) comment only line 21019:06
v13(b) comment lines 210 and 21119:06
HukkaHa, finally19:06
HukkaNo, I ment app side19:06
v13press the button with the label, select something from the list, repeat as needed19:07
HukkaCan't reproduce19:07
HukkaNow19:07
HukkaTakes a long while...19:07
v13i'm having this in n900, scratchbox *and* my debian19:07
v13so it can't be a coincidence19:08
v13did you repeat 5-10 times ?19:08
HukkaMore19:08
HukkaAround 2019:08
v13it crashes in less that 2019:08
*** ppenz has quit IRC19:08
HukkaFirst I tried the same button19:08
HukkaI mean option19:08
HukkaThen I started to vary19:08
v13it doesn't change anything what you select19:08
HukkaNow it took 12 times19:09
v13it always returns a different entry from the previous time19:09
v13it crashed ?19:09
HukkaYes yes19:09
v13see? :)19:09
HukkaHuh? This was the third time19:09
v13there is something i just observed:19:09
HukkaLike I said, last time it took aroun 20 times19:09
HukkaYou just insisted that it should happen faster19:10
v13it is somehow random.19:10
HukkaThen it took 1219:10
v13well in my pc takes exactly 1 with 210 commented19:10
v13but it's a random effect most of the time19:10
v13(with more code)19:10
*** DocScrutinizer51 has quit IRC19:10
v13if you run the program and select something, see the id: XXXX line that is shown19:10
v13it is delayed19:10
v13no idea why19:11
*** Docscrutemp has joined #maemo-devel19:11
*** Docscrutemp is now known as DocScrutinizer5119:11
v13it's the last line that is printed (no need to crash)19:11
v13this is line 5319:12
*** achipa has quit IRC19:14
v13I can save you a lot of time if you tell me what you have in your mind19:18
HukkaI'm grokking it.19:18
HukkaOr trying to19:18
v13"grokking" ?19:18
HukkaHeh19:18
* v13 not english-speaking19:18
HukkaRead Heinlein man, he's a great writer :)19:18
HukkaWell, it's not really English, I think19:19
v13found it in wikipedia :)19:19
Hukkav13: Also, please put spaces around assignment operators and comparisons19:21
HukkaJust saying, it makes it nicer to read for other Python people19:21
v13yeah.. i'll try19:21
Hukka:)19:21
v13but i can't do that easily. C habits.19:21
v13any idea on why the "id:" line takes so much time ?19:22
v13to be shown i mean19:23
HukkaNot there yet, man19:23
HukkaOh, also, class names start with capital letters19:23
v13that's something i usually do19:23
v13just ignore the first two classes :P19:24
HukkaHey, I can't, that's where it prints the id :)19:24
HukkaI think it's this funky way how you implement the whole thing19:27
HukkaI haven't really ever done a GUI, that keeps creating new elements19:28
v13it's not the actual code19:28
HukkaSee http://www.riverbankcomputing.co.uk/static/Docs/PyQt4/pyqt4ref.html#garbage-collection19:28
v13i'm awayre of gc19:28
v13aware19:28
v13that's a portion of the code. After failing with everything else i started removing ilnes19:28
v13lines19:28
v13so it doesn't make much sense19:29
HukkaI think that if you kept a reference to both kinds of layouts and just hid the other, things might work better19:29
v13well... should that happen ?19:29
HukkaNow there's a high chance that Python and C side have different view of what points where and what can be destroyed19:29
HukkaEventually either tries to GC something that isn't actually there anymore19:29
v13i've tested that19:30
v13by adding destructors19:30
v13it didn't happen19:30
v13i also used delayed signals to make sure that this *happens*19:30
v13nothing there too19:30
v13I can try keeping the uis too19:30
v13one mom19:30
HukkaStill, why destroy and recreate things all the time?19:30
v13to save memory19:30
v13why keep all objects ?19:31
v13what if I add 10 such things ?19:31
HukkaStill, it's only ten buttons, or maybe ten times five19:32
HukkaOnly couple of buttons19:32
v13and a graph19:32
v13two graphs19:32
v13and there will be more19:32
v13I believe that the id: line is the key19:33
v13it's a signal that should be delivered a lot earlier19:33
*** onen|openBmap has joined #maemo-devel19:34
v13is there a way to dump a backtrace on python ?19:35
v13I throwed an exception but it stops somewhere19:35
v13traceback19:36
HukkaBut since this isn't really even happening only on maemo, you could ask the pyqt people directly19:36
v13well... true19:37
HukkaThey have way better idea on what funky things results to what19:37
HukkaI'm pyside user anywawy19:37
HukkaJust saying that this approach looks... wrong19:37
v13do you know of a pyqt channel here ?19:37
v13how would you do it w/o keeping objects around ?19:37
HukkaNot sure, never had that kind of case19:38
v13perhaps it is not meant for a window to change its central widget19:38
HukkaYeah19:39
*** SpeedEvil1 has joined #maemo-devel19:39
*** ph1l has quit IRC19:40
v13i'm trying #pyqt... thanks for helping!19:40
v13I'll have pep 8 in mind :)19:40
*** DocScrutinizer51 has quit IRC19:41
*** Docscrutemp has joined #maemo-devel19:41
*** Docscrutemp is now known as DocScrutinizer5119:41
HukkaI'll stick around there and see what they have to say19:41
HukkaNow that I invested this much time, I'd rather see the right answer too19:42
*** SpeedEvil has quit IRC19:42
*** SpeedEvil1 is now known as SpeedEvil19:43
*** dazo is now known as dazo_afk19:44
v13:)19:47
*** SpeedEvil has quit IRC19:55
*** achipa has joined #maemo-devel19:55
*** yerga has quit IRC20:11
*** dazo_afk is now known as dazo20:15
Ian--people i'm trying to find a sample to create a desktop widget in Qt20:16
Ian--any ideas from where i can find something to get me started?20:16
*** lcuk2 has quit IRC20:16
*** lcuk has joined #maemo-devel20:16
v13ian: afaik, qt widgets aren't supported in maemo520:20
v13only as external programs20:20
v13 here: http://wiki.maemo.org/Qt4_Hildon#Home_widget_interaction20:22
v13"Currently Qt Maemo lacks full support for Hildon widgets introduced in Maemo 5."20:22
v13 "There is work ongoing to provide Qt Hildon widgets for Maemo 5, see Qt Hildon Widgets"20:22
VDVsxv13, you can use them from Qt, but in a hackish way20:25
VDVsxaccording to the Qt folks next release will have full support20:25
v13good20:25
Ian--ic so that's y i couldnt find anything20:27
HukkaThe same wikipage says, that 4.6 works20:27
HukkaIan--: Plz uz fl wrds20:28
*** guido_g has joined #maemo-devel20:36
*** guido_g has left #maemo-devel20:36
*** Acedip has quit IRC20:48
*** DocScrutinizer51 has quit IRC20:48
*** Docscrutemp has joined #maemo-devel20:48
*** Docscrutemp is now known as DocScrutinizer5120:48
*** guido_g has joined #maemo-devel20:53
*** guido_g has left #maemo-devel20:53
*** DocScrutinizer51 has quit IRC21:08
*** Docscrutemp has joined #maemo-devel21:08
*** onen|openBmap has quit IRC21:09
*** Docscrutemp is now known as DocScrutinizer5121:09
*** SpeedEvil has joined #maemo-devel21:23
*** slonopotamus has quit IRC22:00
v13Hukka thanks for your help too!22:05
HukkaHuh? Solved it?22:06
* Hukka checks #pygt22:06
HukkaWeren't you supposed to _not_ store the different elements :?22:07
v13well..22:07
v13pyqt was supposed *not* to dump core :P22:08
v13I can write C for that :)22:08
v13wifieye would have at least one more view if that worked at first :)22:09
*** thomaz has joined #maemo-devel22:13
*** onen|openBmap has joined #maemo-devel22:25
*** ph1l has joined #maemo-devel22:26
*** povbot has joined #maemo-devel22:45
*** sleipnir has quit IRC22:51
*** dazo is now known as dazo_afk22:57
*** BabelO has quit IRC23:02
*** thomaz has quit IRC23:04
*** ph1l has quit IRC23:04
*** thomaz has joined #maemo-devel23:05
*** BabelO has joined #maemo-devel23:07
*** jebba has joined #maemo-devel23:13
v13is it possible to make a widget paint on a painter that I'll provide ?23:21
*** ph1l has joined #maemo-devel23:22
*** onen|openBmap has quit IRC23:30
*** ph1l has quit IRC23:31
*** kinta has joined #maemo-devel23:35
HukkaCustom painting?23:43
HukkaShould be trivial23:43
HukkaSee examples from the qt blog, for example23:43
v13hm.23:55
v13I'm trying to do this but i've stopped on something else.23:55
v13I've a model and a listview. Model gets data dynamically. is there a function that asks for listview to update? repaint doesn't work because it doesn't ask for the number of rows()23:56
v13i.o.w. who are qlistview's data refreshed ?23:57
v13i'm now looking in begininsertrows() but it seems to complex23:59
v13beginInsertRows()23:59
« Wednesday, 2010-02-17 Index Friday, 2010-02-19 »

Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!