**** BEGIN LOGGING AT Thu Mar 17 02:59:57 2011 Mar 17 03:20:29 Since duicontrolpanel's wifi applet is DOA... anybody know how to use connman from cli? Mar 17 03:24:03 gabrbedd: install connman-test, and use the tools installed. Mar 17 03:24:56 TSCHAKeee: thx... Mar 17 03:25:13 * gabrbedd goes find a flash drive since it's a network catch-22 Mar 17 03:26:57 can't you build a new image? Mar 17 03:27:12 just shove it into the kickstart Mar 17 03:29:04 TSCHAKeee: sure... if I want to wait 30-60 mins for a new image. :-) Mar 17 03:33:17 \o/ Mar 17 03:33:26 * gabrbedd hugs TSCHAKeee Mar 17 03:33:52 yw Mar 17 03:33:55 :) Mar 17 04:00:34 * gabrbedd is rsync'ing a kernel build tree over wifi Mar 17 05:04:25 HI I m trying to buld rpm-python from meego 1.1 release area using OBS. the spec file has no License feild , the build fails giving Leicense error, also the spec file is name rpm-python.spec but the spec fil is looking for rpm.spec, is it a bug? Mar 17 06:39:28 lbt: ping Mar 17 06:40:36 lbt: the armv8el scheduler is enabled in cobs but the build needs to be updated to newer version the current one that is installed there doesn't work well with armv7hl. build version 2011.01.20 and newer works. Mar 17 08:02:24 hmmh. Do I need to register for the conference when I've submitted a session proposal or should I wait if it get approved since the speaker registration requires some code? Mar 17 08:02:56 ask on meego-events@ or -community@ Mar 17 08:03:23 will do. thanks Mar 17 08:07:52 looks like it's been answered already: http://lists.meego.com/pipermail/meego-events/2011-March/000058.html Mar 17 08:50:42 morning \o Mar 17 09:01:07 lcuk: morning Mar 17 09:01:19 i'm pretty close to pushing my mxr changes to upstream Mar 17 09:02:11 * timeless_w7ip goes to try to merge them to mxr.meego.com Mar 17 10:02:20 hi All, Currently I have Meego1.1 set up with QEMU Mar 17 10:02:36 i want to upgrade it to 1.2 prelreleases like 1.1.90 or something Mar 17 10:02:52 could someone provide some input on this Mar 17 10:09:06 Sage: is there a bug logged? Mar 17 10:12:29 good morning Mar 17 10:14:10 morning venemo \o Mar 17 10:15:07 hey lcukn900 :) Mar 17 10:15:07 Stskeeps: did we have a build update requirement bug for hardfp armv8el scheduler at some point? Mar 17 10:15:11 what's up? Mar 17 10:15:19 Sage: build 2011.01.20 or .30 Mar 17 10:15:57 I think .20 was enough for armv8el Mar 17 10:16:46 at least my local obs is running build-2011.01.20-47.2.noarch and that works Mar 17 10:17:02 it would be nice how smoother will puzzle-master run on hardfp... since Qt Graphics View uses lots of fp calculations Mar 17 10:17:21 lcukn900: could you test it someday? Mar 17 10:17:49 Stskeeps: ^ Mar 17 10:18:28 Venemo_N900: when I get my other N900s' battery back in a week I'll test it for you ;) Mar 17 10:18:38 thanks alterego :) Mar 17 10:18:42 venemo, no - you can test it yourself! Mar 17 10:18:47 Heh Mar 17 10:19:04 I can't, I have no sd card Mar 17 10:20:20 it is ok, we have no general harpfp build yet, so whilst that is being sorted out you can get an sd card :p Mar 17 10:20:52 Heh Mar 17 10:22:09 Stskeeps: do you have a blog? Mar 17 10:22:14 i want to upgrade it to 1.2 prelreleases like 1.1.90 or something ....could someone provide some input on this Mar 17 10:22:50 lcukn900: I can't spare the price of an sd card now. Mar 17 10:22:54 thiago: only mer-project.blogspot.com but i don't write too much these days Mar 17 10:23:01 Stskeeps: ok, n/p Mar 17 10:23:22 bkalinga update versions is complex and not tested for, perhaps you could just get a 1.1.90 image Mar 17 10:23:30 some might claim #meego-arm logs is my microblog ;) Mar 17 10:24:23 if they did it is like having a blog in the basement behing a sign saying beware of the tiger :p Mar 17 10:24:33 Stskeeps: :-) Mar 17 10:24:57 Tiger Blood Mar 17 10:25:02 1.1.90 image for QEMU how would i get it..can you share any pointer Mar 17 10:25:09 alterego: :) Mar 17 10:26:31 * arfoll is all happy with his wetab :-) Mar 17 10:26:32 lcukn900: currently i don't have a device and trying out on QEMU...so don't know how to get 1.1.90 image Mar 17 10:28:47 bkalinga: searching the internet for "meego 1.1.90 qemu" somehow strangely gave me a link to http://repo.meego.com/MeeGo/builds/1.1.90/1.1.90.1.20110201.1/handset/images/meego-handset-ia32-qemu/ Mar 17 10:28:53 * lcukn900 shocked Mar 17 10:29:21 let me check that ink..thanks lcukn900: Mar 17 10:29:27 it might not be latest latest one but the folder tree should point you in right direction Mar 17 10:30:36 any idea how i can update this location in my qemu so that zypper will pick the correct source rpm Mar 17 10:30:48 and i will build them manually Mar 17 10:37:56 lcukn900: you need to use Mar 17 10:47:56 lcukn900: the link u shared contains a meego-handset-ia32-qemu-1.1.90.1.20110201.1-raw.tar.bz2 Mar 17 10:48:14 i am not getting what should i do further Mar 17 10:49:36 bkalinga: might you try google? Mar 17 10:58:02 Sage: updated build Mar 17 11:08:02 timeless_w7ip: actually googled it but not able to find anything related to "how to upgrade meego 1.1 to 1.1.90 " Mar 17 11:08:20 if you have any link ..would you please share it Mar 17 11:18:04 bkalinga: what you want to is to get a new target Mar 17 11:18:15 http://wiki.meego.com/SDK/Docs/1.1/Getting_started_with_the_MeeGo_SDK_for_Linux#Install_MeeGo_Target Mar 17 11:19:09 Don't google as your first move, go first to the relevant documentation Mar 17 11:19:23 What you need might actually be there :) Mar 17 11:25:56 grr.. i don't like apache Mar 17 11:37:52 timeless_w7ip: maybe use httpdlight instead? Mar 17 11:38:16 cosmohill: so... Mar 17 11:38:26 i'm writing an app that lives behind a web server Mar 17 11:38:33 and i was trying to accelerate it by talking to apache Mar 17 12:10:37 jonnor: This is the link i had and i already followed to install Meego1.1 Mar 17 12:11:10 It does not say anything about Meego1.1.90 installation or how to upgrade? Mar 17 12:12:50 if i follow this link it seems i need to add http://repo.meego.com/MeeGo/builds/1.1.90/1.1.90.1.20110201.1/handset/images/meego-handset-ia32-qemu/ to /etc/apt/sources.list.d/meego-sdk.list Mar 17 12:13:19 then next step is "Add the repository public key " Mar 17 12:13:54 would it be the same as in the link for Meego1.1 i am not sure..any idea Mar 17 12:15:42 bkalinga: 1.1.90 is just another target Mar 17 12:16:23 the repo key is probably the same Mar 17 12:16:33 and if you allready added it then you can skip that step, ofc Mar 17 12:16:52 in any case, you can probably just get the tarball for the target yourself Mar 17 12:17:16 ok if i manually get the tar then Mar 17 12:17:24 what next i need to do Mar 17 12:17:52 how to add it to the "mad-admin list " Mar 17 12:18:19 once it will reflect in mad-admin list i can use that..if i am correct Mar 17 12:19:39 i think i can place that that .tar to a specified path and run sudo mad-admin create -f to create a target out of that Mar 17 12:20:51 any idea where i need to place the tar after downloading and what must be the value for to the command create -f Mar 17 12:26:11 bkalinga: man mad-admin Mar 17 12:26:24 if that fails, locate your existing targets in the filesystem Mar 17 12:26:49 ok...let me try Mar 17 12:27:15 I suspect you also want a runtime http://wiki.meego.com/SDK/Docs/1.1/Configuring_QEMU_runtimes Mar 17 12:31:55 http://pastebin.com/hEHiMrP9 Mar 17 12:32:12 i have all this from sudo mad list Mar 17 12:32:26 but only two targets are installed Mar 17 12:32:59 there No manual entry for mad Mar 17 12:34:48 me neeither prob didnt install the doc Mar 17 12:41:00 Uk ne1 Mar 17 12:43:14 erh Mar 17 15:02:57 * timeless_office can't find eric Mar 17 15:03:59 have you looked behind the fridge? Mar 17 15:07:24 or under it Mar 17 15:16:01 * timeless_office offers ask a phone Mar 17 15:17:42 timeless_office, which eric are you looking for? Mar 17 15:18:12 the one who can poke bugzilla to give me a component/product Mar 17 15:23:34 timeless_office: I thought it was bdub Mar 17 15:30:18 gabrbedd: i dunno Mar 17 15:30:22 dawnfoster? Mar 17 15:30:52 hey timeless_office Mar 17 15:30:56 gabrbedd: if you find any pokable someone who you think can help me get my bugzilla product/component, please feel free to poke them for me Mar 17 15:31:16 dawn: hey, so... i asked for a bugzilla component a while ago, can you see about poking someone to make it happen? :) Mar 17 15:32:27 I'm still catching up - let me look into it Mar 17 15:32:34 bug 11514 Mar 17 15:32:37 <_MeeeGoBot_> Bug https://bugs.meego.com/show_bug.cgi?id=11514 enh, Undecided, ---, eric.le-roux, NEW, please add a product mxr.meego.com in MeeGo Community Infrastructure classification Mar 17 15:32:45 it was filed in december.. Mar 17 15:33:10 anyway, at this point 2/3 of the roots are automatically updating Mar 17 15:33:43 the last one is close to the point where i can let it loose, and i'm close to the point where i think i can release my elevation principles Mar 17 15:33:54 or at least start thinking about letting someone audit the code Mar 17 15:44:49 dawn: oh, at some point we need to remind people to take ownership of bugs when they resolve them Mar 17 15:45:04 the QA of one of the bugs in that component resolved a bug but didn't claim ownership Mar 17 16:14:13 Amazing: A decade of GStreamer hacking activities squeezed into ~5 minutes. http://www.youtube.com/watch?v=VMiMGOg53Ec Makes me feel so proud of being part of a creative community! Mar 17 16:15:32 qgil: Yeah, that visualization tool is cool. A few years old though. Mar 17 16:16:05 X-Fade: first time i see it Mar 17 16:16:15 me too - very cool Mar 17 16:16:28 qgil: Ah, no. I have seen it for many projects over the last few years. Mar 17 16:17:49 But still, really cool to see. Mar 17 16:18:34 X-Fade lbt I'm assuming you will have one conference proposal for apps.meego.com, right? Even one separate from the other one expected about the community OBS? Mar 17 16:19:00 qgil: Do you think that is needed for the target audience? Mar 17 16:19:18 of course! nobody knows anything about apps.meego.com today Mar 17 16:19:21 I didn't think it was the kind of thing wanted Mar 17 16:19:37 well, good that I checked then Mar 17 16:19:39 I was doing more of a 'vendor' talk Mar 17 16:19:42 qgil: Then the conference really gives off the wrong impression. Mar 17 16:19:42 yep Mar 17 16:19:50 *conference site. Mar 17 16:19:59 ? Mar 17 16:20:20 Well, more the vendor/commercial counter part this time. Mar 17 16:20:45 I'm not saying that this should be a keynote - but of course any relevant novelty in the MeeGo project has and will always have a place in a MeeGo Conference Mar 17 16:20:56 But yes, I can come up with a proposal for apps.meego.com and how a vendor can benefit from community apps. Mar 17 16:21:15 by tomorrow Mar 17 16:21:44 Sure, tomorrow is a long day :) Mar 17 16:21:54 And we still have today. Mar 17 16:22:06 just curious: what exactly are you seeing in the website hinting to this vendorYES-communityNO focus? Mar 17 16:22:32 a site with a t-shirt contest promoted in the home, funny characters... why is this different from Dublin? Mar 17 16:23:33 qgil: That was the general impression when talked about the target audience on the list and meeting? Mar 17 16:23:55 * lbt agrees - but the site doesn't give that impression Mar 17 16:24:13 qgil: That's the impression I get, too... so it makes me hesitant to do something too technical. :-) Mar 17 16:24:25 No, sorry that is me wording it the wrong way. Mar 17 16:24:32 There is indeed an intention to actively bring commercial players... but I always assumed that the community basics are in place Mar 17 16:24:47 aside: "Harmattan in the MeeGo community" ... :) Mar 17 16:25:52 This is interesting: EVERYBODY HERE - PLEASE DON'T STOP YOURSELVES FROM SUBMITTING A PROPOSAL TO THE MEEGO CONFERENCE Mar 17 16:25:56 lbt: Well, the "about" page says the conference is a showcase for new meego products. Mar 17 16:26:32 hi all... nice to meet this channel !! X:) Mar 17 16:26:47 * w00t_ has no idea what he'd propose Mar 17 16:26:51 very cool video Mar 17 16:27:41 "From application developers to platform designers, hardware vendors to open source enthusiasts, anyone engaged in MeeGo is invited to contribute and participate at the San Francisco Regency Hyatt, California, USA, from May 23rd-25th, 2011." Mar 17 16:27:44 gabrbedd: true ... but that doesn't exclude other things. I did somehow get the feeling that the conferences would alternate between "less community" and "more community" oriented Mar 17 16:27:53 http://sf2011.meego.com/program/call-session-proposals Mar 17 16:28:35 So yes, the summary of the conference tries to attract commercial vendors, but I believe there has never been an attempt to push back the community, hobbyists, individuals on their own etc Mar 17 16:28:37 can I change my Symbian^3 OS to Meego? Mar 17 16:28:51 vldmr: short answer? no. Mar 17 16:29:17 Like, if Dubling was 1000 people all of them could repeat and have the extra 400 seats for ISV and device vendors with a presence in the Bay Area / US Mar 17 16:29:43 TSCHAKeee: ok... and can I change Maemo to Meego ? Mar 17 16:30:13 vldmr: there's a lot of effort to make meego relevant to maemo users Mar 17 16:30:13 vldmr: i'm sorry but your questions are not easy to answer Mar 17 16:30:17 vldmr: again, if you're asking this question in this manner...it's not ready for you. Mar 17 16:30:28 lbt: Also, I don't know what I'm doing when it comes to conference presentations. :-) Mar 17 16:30:43 gabrbedd: talking to friends Mar 17 16:30:57 lbt: ...so it's hard to break away from the code to make a powerpoint. :-) Mar 17 16:31:10 impress please Mar 17 16:31:18 or html5 Mar 17 16:31:38 or S5 Mar 17 16:32:07 I would like to use Meego in my Nokia device !! can I do it anyway ? Mar 17 16:32:17 lbt: yeah, yeah... sorry, everyone around me lways says "powerpoint" Mar 17 16:32:19 what device vldmr Mar 17 16:32:32 gabrbedd: chip away at them Mar 17 16:32:52 lbt: Nokia C6-00 Mar 17 16:33:01 gabrbedd: it's like a performance for people who can't sing Mar 17 16:33:11 lbt: smartphone Mar 17 16:33:27 vldmr: no.... and to be fair ... it's N900 or nothing Mar 17 16:33:30 vldmr: that phone is much too small for MeeGo. Mar 17 16:33:52 Nokia have made some statements about a MeeGo device coming out ... google should help Mar 17 16:34:26 berndhs: karaoke night? Mar 17 16:34:33 :-p Mar 17 16:34:41 gabrbedd: or mime if you don't like your voice Mar 17 16:35:59 * gabrbedd is trapped in a shrinking, invisible box Mar 17 16:37:56 lbt: gotta choose my battles. Besides, I don't use PowerPoint *or* Impress. I prefer a chalk board. :-) Mar 17 16:39:05 lbt: anyway, I'll see if I can dream up an idea over lunch. I'm already teaming with texrat for one, btw Mar 17 16:39:22 dont they just hook up your laptop to a projector? so you can run anything you want ? Mar 17 16:39:43 * lbt wonders how far it is to Whistler from SF.... Mar 17 16:39:56 lbt: far Mar 17 16:40:05 shh Mar 17 16:40:05 lbt: Spring skiing? :) Mar 17 16:40:15 have you *seen* the reports Mar 17 16:40:25 vldmr: the C6-00 is not even S^3. Anyway, the problem is not the OS, it's the HW. If you know how to do it, you can. But I'm telling you that it's not easy to flash a different OS on that HW Mar 17 16:40:39 Stskeeps: are you ok with extending? Mar 17 16:41:09 berndhs: i like my chalk *very* soft, and *very* dusty. :-) Mar 17 16:42:10 berndhs: yep Mar 17 16:56:36 * lcuk lobs a chalk duster @ gabrbedd Mar 17 16:57:14 merp Mar 17 16:57:57 Sister for sale! Sister for sale! One crying and spying young sister for sale! Mar 17 16:58:12 I'm really not kidding! So who'll start the bidding? Mar 17 16:59:42 lbt, European X-Games just started in France Mar 17 17:18:00 pupnik: I'll see you one sister and raise you two nephews and a neice.... Mar 17 17:20:04 shel silverstein was a good children's poem writer lbt Mar 17 17:20:47 * lbt likes your timestamp Mar 17 17:22:50 anyone here know how I get Aegis to allow my to install my own deb files over the ones from com.nokia.maemo? Mar 17 17:24:15 mathias_m: I'm pretty sure nobody will answer you because you talk of something which is not available yet :) Mar 17 17:24:53 oh... :-/ maybe you can answer off-line? Mar 17 17:25:07 please tell us more :) Mar 17 17:25:09 mathias_m: No, don't look at me. Just sayin' Mar 17 17:25:40 I know naa-thing... Mar 17 17:25:44 mathias_m: I *think* (from what I remember from Amsterdam), your package needs to come from a domain with a higher trust level; and certain privileges granted to it, than the com.nokia.maemo packages Mar 17 17:25:55 s/domain/repo/ Mar 17 17:25:55 Jaffa meant: mathias_m: I *think* (from what I remember from Amsterdam), your package needs to come from a repo with a higher trust level; and certain privileges granted to it, than the com.nokia.maemo packages Mar 17 17:26:32 *sigh* Mar 17 17:26:56 dpkg -i doesn't work? Mar 17 17:27:00 I have a fix for a problem qt issue and I just want to test it on harmattan hardware Mar 17 17:27:11 have build the binaries, but can't install them Mar 17 17:27:14 w00t_: Sorry I missed your ping - I had *just* left for the day. I just sent a response to your email. Mar 17 17:27:21 cheers Mar 17 17:27:27 kaitlin__: sure, np, good that we're on the same page Mar 17 17:27:41 mathias_m: Send the hardware to me and I'll test it for you. I won't be able to succeed, but at least I'll have a new toy Mar 17 17:27:42 I'll see if I can find some time to start work on it in 30-60 mins after I've wrapped things up at the 'day job' Mar 17 17:27:43 kaitlin__: sorry, was hoping you were talking to me... ;) Mar 17 17:28:15 mathias_m: ;) Mar 17 17:29:13 w00t_: That would be great - thanks so much for the code review. I'm pretty new to Qt development, and the async code was one of the first items I worked on back in Dec. It's been sitting around since then. And it's not well tested. Mar 17 17:29:30 w00t_: So it's great to get a code review and have another set of eyes taking a look. Mar 17 17:29:36 :) Mar 17 17:29:37 happy to help Mar 17 17:30:09 w00t_: Would it be easier if I reviewed / applied your patched beforehand? I don't want you to have to rebase too often. Mar 17 17:30:29 kaitlin__: if you have the time to do that, it would be great, but they don't really touch seasidesyncmodel Mar 17 17:30:33 so I think we'll be ok for the time being Mar 17 17:31:30 * kaitlin__ looks. Mar 17 17:32:19 Ohh.. right, they only touch the listmodel. My understanding from Connie was that there was some interest in having a list model, but then no one really used it. Mar 17 17:32:51 So I think that code is a bit forlorn indeed. Mar 17 17:35:14 kaitlin__: hmmm, if it's not used at all - perhaps it can just be nuked from orbit :) Mar 17 17:36:41 I want to inherit everyone's thinking process in here Mar 17 17:37:32 ray24, *everyones*? are you sure Mar 17 17:37:39 ye Mar 17 17:37:41 s Mar 17 17:37:46 thiago: yes Mar 17 17:37:51 I want to speak like everyone in here. the competent Mar 17 17:38:29 w00t_: Just double checked with Connie, and it does seem like no one is using it. Do you have a need for the listmodel? Mar 17 17:39:13 kaitlin__: no, I'm just generally interested in trying to improve the middleware stack, and libseaside etc caught my interest since I have a fair amount of professional knowledge/experience with PIM/contacts area of QtMobility Mar 17 17:39:28 *waves* ^ yeah that code should be ripped out. it's changed a few hands before bit-rotting in the repos Mar 17 17:40:34 w00t_: Cool - just wanted to make sure we didn't nuke something that someone wanted to use. Mar 17 17:41:10 mnementh: ah! didn't know you were on IRC too, hi Mar 17 17:41:37 helo Mar 17 17:42:06 help Mar 17 17:42:30 kaitlin__: anyway, to make it clear: I'm fine with that code being removed and dropping those patches :) Mar 17 17:44:22 Excellent. I'll bypass on your patches then - I can work on ripping it out instead. Mar 17 17:44:27 official version of opera mobile for meego apparently confirmed Mar 17 17:44:51 oh? Mar 17 17:44:52 url? Mar 17 17:45:05 and which arch? Mar 17 17:45:32 http://reviews.cnet.com/8301-12261_7-20044304-10356022.html Mar 17 17:46:01 w00t_: hi! originally the seperate models came from different apps needing different but shared views... then the apps fluctuated from MTF to QML and libseaside sort of sat there. thanks for your code input though. Mar 17 17:46:29 i'd guess it's more for mobile phones so my guess would be arm Mar 17 17:46:44 mnementh: is seaside still used in the QML stack? Mar 17 17:47:16 mnementh: i'd have figured that it would be a useful place to e.g. do aggregation from multiple QContactManager backends Mar 17 17:49:00 actually, this is much more official source: http://www.opera.com/press/releases/2011/03/17_2/ Mar 17 17:52:55 w00t_: a re-written version is bundled with the QML contacts app since I needed Q_INVOKABLE methods for QML.... i fluctuated between writing it all in the qt-declarativecontacts ContactModel but with some apps still in MTF I had to keep the models in cpp. Which gives us more control for now anyways. We thought of axing libseaside entirely except for the note above about MTF. I'm currently playing around with libfolks to handle aggregation... we'll see how Mar 17 17:54:00 mnementh: ah, folks - nifty, presumably you have a QContactManager interface bolted onto it so QML can talk to it? Mar 17 17:54:38 that's a good approach, anyway Mar 17 17:54:49 so long as there's one central place where aggregation is done, it's great :) Mar 17 17:55:07 * w00t_ looks forward to seeing this stuff so he can tinker with it Mar 17 18:07:49 * CosmoHill would glare at the stuck Windows logo but this graphics card is having other ideas Mar 17 18:16:34 kaitlin__: something else I just realised, good for future reference.. by adding the requests to SeasideSyncModel directly (and not the private member), there was an ABI break (the size of SeasideSyncModel would have changed) which is usually not a good thing in shared libraries Mar 17 18:16:39 so I'll fix that at the same time Mar 17 18:18:04 w00t_, that ABI problem has been the crux of many C++ issues Mar 17 18:18:35 lcuk: it's not a C++ issue, C has the same problem Mar 17 18:19:45 w00t_, in C you do not generally create subclasses for structs Mar 17 18:19:58 lcuk: you do, if you want to avoid this problem Mar 17 18:20:41 (either that, or you just don't add data members, or you announce that you're breaking ABI and bump soversion) Mar 17 18:22:22 lcuk: If you add a new data member to a public class or struct in C or C++, then you will break binary compatability (but not source compatability) Mar 17 18:25:40 gabrbedd, it is entirely feasible to add new members to structures in the library and have no code breakage and apps continue to work happily Mar 17 18:26:15 if the apps have their own instance members then you get breakage Mar 17 18:27:23 * thiago_home recommends reading his very detailed "Binary Compatibility Issues With C++" page Mar 17 18:27:41 thiago_home, links help! (I have read it in the past) Mar 17 18:28:03 google for "binary compatibility c++", it should be the first link Mar 17 18:28:05 lcuk: But if you add it as a member in the private struct (an opaque pointer)... then you don't have to worry about it. :-) Mar 17 18:28:20 http://techbase.kde.org/Policies/Binary_Compatibility_Issues_With_C++ Mar 17 18:28:25 how do you spell compatability? Mar 17 18:28:27 :-p Mar 17 18:28:40 the same way I do my coding Mar 17 18:28:42 copy and paste :-) Mar 17 18:29:12 w00t_: hahaha! Mar 17 18:29:15 many legal systems do the same, its called "precedence" :) Mar 17 18:29:41 "compatible" -le +ility Mar 17 18:30:07 o/ thiago_home Mar 17 18:30:16 good day? :) Mar 17 18:30:18 yes Mar 17 18:30:44 I think this is easier in Italian, where the consonants haven't changed Mar 17 18:30:46 day @ home = good day Mar 17 18:30:48 mine wasn't, I spent pretty much all of it fighting with a stupid login dialog Mar 17 18:31:14 thiago_home: Thanks for the article! :-) Mar 17 18:31:19 w00t_: that means no work done if you couldn't get past the login dialog Mar 17 18:31:38 exactly Mar 17 18:31:46 the bug that I was trying to work on needed the dialog to work Mar 17 18:31:50 and it didn't, for the most part Mar 17 18:32:08 * thiago_home last updated that page when he found this issue: http://techbase.kde.org/Policies/Binary_Compatibility_Examples#Override_a_virtual_that_doesn.27t_come_from_a_primary_base Mar 17 18:33:26 how does compatibility in java or c# or .net work? Mar 17 18:33:51 I don't know for sure Mar 17 18:34:15 when I last looked into Java bytecode, I think the function names were stored in the .class files Mar 17 18:34:22 so it does lookup-by-name Mar 17 18:34:27 It does Mar 17 18:34:31 And signature Mar 17 18:34:38 do developers have to concern themselves about it? Mar 17 18:34:51 since all objects are held as pointers with refcounting, the size changing is not an issue Mar 17 18:34:54 I recall from old vb days there were compatibility gotchas Mar 17 18:35:10 I don't know about member layout changing, but I would expect that not exist either Mar 17 18:35:15 lcuk: A little, but it's rare to change a JAR in a Java system without rebuilding and recompiling the rest of the app Mar 17 18:35:24 i.e., treat members as properties and make calls to get / set Mar 17 18:35:28 thiago_home, I thought that was how C++ did its objects tbh Mar 17 18:35:45 lcuk: Object *ptr = new Object; <--- this is a pointer Mar 17 18:35:51 I think you can even change a signature to be more acceptable and you still don't need to recompile. Mar 17 18:35:54 Jaffa, so a core update comes in and all your java apps carry on working? Mar 17 18:35:56 lcuk: Object ptr; <--- this is by value Mar 17 18:36:14 the compiler reserves sizeof(Object) in the stack or in the structure, then calls the constructor Mar 17 18:36:20 e.g. library changes foo(Blah) to foo(Object) and anything will continue to work without recompile. Mar 17 18:36:28 thiago_home, yes that is where subclassing screws up on c++ Mar 17 18:36:35 lcuk: Yes. Thinking about it we do it a lot at work. Mar 17 18:36:38 because it stores a byval instance member of the baseclass at the head Mar 17 18:36:44 lcuk: not only subclasses, but just using the class too Mar 17 18:37:00 thiago_home, yeah, that is stack based bashing usually Mar 17 18:37:01 lcuk: I've had that often when a program randomly crashes and not even the debugger and disassembly make sense. Mar 17 18:37:19 well, it looks like my session proposal to the MeeGo Conference is being useful advertising the event.................. Mar 17 18:37:21 "when a crash doesn't make sense, recompile" Mar 17 18:37:35 happens often to qmake Mar 17 18:37:37 lcuk: Things which make it easy in Java: (almost) everything is virtual, binding is done by name, type & signature as runtime linking and every object is passed by ref. Mar 17 18:37:55 Jaffa, nice indeed Mar 17 18:38:07 * timeless_office sighs Mar 17 18:38:08 * thiago_home points out that the last time he looked at java bytecode was at least 12 years ago, probably more Mar 17 18:38:13 qgil, I saw that, and people are blogging and stuff about Harmattan using your proposal Mar 17 18:38:14 qgil: i saw it made engadget ;) Mar 17 18:38:20 lcuk: want to learn some basic bits about c++? :) Mar 17 18:38:31 timeless_office, just discussing Mar 17 18:38:33 if it made to engadget, it must be true! Mar 17 18:38:52 I did C++ at college years ago Mar 17 18:38:56 * lcuk actually qualified in it Mar 17 18:38:58 lcuk: I don't think subclassing really has much to do with it. Mar 17 18:39:07 thiago_home: The JLS folks go to great lengths for forward (and, as much as possible) backwards compatibility. Hence the "interesting" approaches to generic, annotations and enums in Java 5 Mar 17 18:39:15 gabrbedd, but that is where an instance member is stored for the base class Mar 17 18:39:21 and if the base class expands Mar 17 18:39:46 Jaffa: the same can't be said about the C++ committee members... Mar 17 18:39:51 frals? A session proposal made Engadget?! No offence to qgil, but it wasn't that earth-shattering ;-) Mar 17 18:40:00 Jaffa: ABI is a completely fobidden topic Mar 17 18:40:00 lcuk: But that jacks up everything... not just the derived class. Mar 17 18:40:13 I think it's the second time I make it to Engadget - the first one was thanks to some post #127 in a talk.maemo.org about enabling dual-booting in the N900 Mar 17 18:40:42 Jaffa: yeah, i dont see what the noise is about either, guess its a slow rumour day ;) Mar 17 18:40:51 * thiago_home is hoping that the new lvalue-this member functions will be ABI compatible with the current regular member functions Mar 17 18:41:02 lcuk; bug 14660 / bug 14661 Mar 17 18:41:04 <_MeeeGoBot_> Bug https://bugs.meego.com/show_bug.cgi?id=14660 nor, Undecided, ---, peter.j.zhu, NEW, icu gencfu uses delete instead of delete[] for memory allocated by new[] in readFile Mar 17 18:41:06 <_MeeeGoBot_> Bug https://bugs.meego.com/show_bug.cgi?id=14661 nor, Undecided, ---, peter.j.zhu, NEW, exempi CDFileHeader::release() uses delete instead of delete[] for new[] from CDFileHeader::read Mar 17 18:41:11 thiago_home: Crazy, especially since system-wide shared libs are more common in C/C++ Mar 17 18:41:18 frals: URL? Mar 17 18:41:26 Jaffa: http://www.engadget.com/2011/03/17/nokia-talking-homegrown-harmattan-release-of-meego-at-may-conf/ Mar 17 18:41:32 timeless_office, eek not removing the arrays Mar 17 18:41:48 frals: The URL is telling enouhgh Mar 17 18:41:49 ok, guys, I have to reboot Mar 17 18:41:59 my 2.6.37 kernel decided to stop talking to my HD again Mar 17 18:42:04 \o thiago_home thanks for the links it will give many food for thought again Mar 17 18:42:23 Jaffa: the thing started with MyNokiaBlog, where the guy made a weird connection of things - in that "tech blogger" way Mar 17 18:44:10 * timeless_office goes for a different class Mar 17 18:44:16 kicked off webchat - interesting Mar 17 18:44:41 quimgil: oh, you too? Mar 17 18:44:46 i guess the service hickupped Mar 17 18:44:50 * timeless_office is glad to be able to ghost Mar 17 18:45:06 lcuk: For example... suppose someone changes struct timeval { time_t tv_sec; suseconds_t tv_usec; } by changing time_t from (e.g.) uint32_t to uint64_t. Mar 17 18:45:43 lcuk: That changes the size of the struct... and the offset of tv_usec and the size of tv_sec. Mar 17 18:45:46 gabrbedd, of course, using struct timeval would because world + dog just use them on the stack Mar 17 18:46:04 if there was a timeval_create() function though that you stored into a pointer Mar 17 18:46:09 the problem would not occur Mar 17 18:46:30 lcuk; people are really very bad at using the right free() function tm Mar 17 18:46:43 in theory C++'s new/delete were supposed to solve this Mar 17 18:47:04 lcuk: Right... and that would be an opaque pointer (like the "private class" that w00t_ was talking about). Mar 17 18:47:07 of course, the existence of new/new[] - delete/delete[] shows that even the basics ... Mar 17 18:47:36 gabrbedd, sure, but in c++ it is normal to have the dual layer indirection Mar 17 18:47:36 and then there's that minor detail that you can't really use new/delete if you're talking to an object from a different compiler Mar 17 18:47:45 which was why I was asking how other languages handled it Mar 17 18:48:56 http://mxr.meego.com/repo.meego.com/source/gnupg/g10/parse-packet.c?mark=1149,1165-1168,1168#1148 Mar 17 18:49:00 anyone see anything wrong there? Mar 17 18:49:28 if mxr ever loads, i'll tell you :P Mar 17 18:50:35 * timeless_office nods Mar 17 18:50:44 i need to look into giving it some help Mar 17 18:50:47 yeah some bozo isn't using brackets to make the meaning clear Mar 17 18:50:50 the database it deals w/ is rather large Mar 17 18:51:15 w00t_: i presume it has loaded by now Mar 17 18:51:41 lcuk: BTW, yesterday there was some discussion of tic-tac-toe. In the US, when there's a tie, was say it's the "Cat's game." Do they do that over there? Mar 17 18:51:47 timeless_office: I think something is screwy with my connection, it's not working Mar 17 18:51:51 it took me 12s to load (a bit long, but oh well) Mar 17 18:51:53 brb anyway, dinner :) Mar 17 18:52:14 thiago, url of interest: http://mxr.meego.com/repo.meego.com/source/gnupg/g10/parse-packet.c?mark=1149,1165-1168,1168#1148 Mar 17 18:52:47 gabrbedd, no, just save and carry on with new game Mar 17 18:52:58 timeless_office: in the green hightlighted section, the expression should be bracketed () properly Mar 17 18:53:09 especially around the macro Mar 17 18:53:27 lcuk: k, thanks. :-) Mar 17 18:53:47 berndhs: that's of course rather annoying, but ignore it, there's something more exciting further into that line Mar 17 18:54:41 well the comment says "return soem value different from NULL" and then it returns NULL in some cases Mar 17 18:55:41 what if it's just 0 and not NULL ;P Mar 17 18:56:05 berndhs: yeah, it's a bit strange, ignore the null bit Mar 17 18:56:54 also returns a pointer onto the parameter stack Mar 17 18:57:24 assuming parameters are on a stack Mar 17 18:57:29 * timeless_office gives berndhs a cookie Mar 17 18:57:37 http://mxr.meego.com/repo.meego.com/source/sudo/sudo.c?mark=692,699,711#685 Mar 17 18:57:49 maligor: 0 is NULL Mar 17 18:58:03 maligor: but there's no guarantee that 0 / NULL is actually saved as a bitwise zero Mar 17 18:58:27 thiago: i love that feature Mar 17 18:58:46 thiago_home, I rather meant that it's logic returning 0 rather than NULL :) Mar 17 18:59:16 int Class:: *pmm = 0; <--- this doesn't save a zero Mar 17 18:59:54 $ g++ -S -o - -xc++ - <<<'class Class; int Class:: *pmm = 0;' | grep -A1 pmm: Mar 17 18:59:57 pmm: .long -1 Mar 17 19:02:20 WM7: Kill It With Fire Mar 17 19:02:40 timeless_office, "if(!pkt_buf)" is one of the first things that gets you into the block in the first place Mar 17 19:02:46 so you already have a 0 Mar 17 19:02:51 maligor =~ s/it's/its/ Mar 17 19:03:22 mind you, it takes the & of it Mar 17 19:03:40