=== salem_ is now known as _salem | ||
=== chihchun_afk is now known as chihchun | ||
=== chihchun is now known as chihchun_afk | ||
dholbach | good morning | 06:46 |
---|---|---|
zbenjamin | akiva-thinkpad: well our next steps now will be to release a opt in package of it in the SDK ppa | 06:56 |
zbenjamin | akiva-thinkpad: and then we'll see whats required next | 06:56 |
akiva-thinkpad | zbenjamin: ah hey | 06:56 |
akiva-thinkpad | bout to head out | 06:56 |
akiva-thinkpad | zbenjamin: how do you want to do the opt in? | 06:56 |
akiva-thinkpad | Just have it in the "About Plugins", but disabled by default? | 06:57 |
zbenjamin | akiva-thinkpad: but i'm afraid if you want to upstream it very soon you would need to do that yourself since we have no timeslot for that atm. Before we can start to upstream our stuff we first wold need the ui toolkit to be upstreamed and then move on from there. But I do not see that we will have enought time soon | 06:57 |
zbenjamin | akiva-thinkpad: no , let people apt-get install qtcreator-plugin-autopilot if they want it. | 06:58 |
akiva-thinkpad | zbenjamin: that's fine. This isn't exactly a priority. | 06:58 |
zbenjamin | akiva-thinkpad: we will not depend on it for now | 06:58 |
zbenjamin | when we are sure its finished and tested enough then it will become a part of the ubuntu sdk | 06:59 |
akiva-thinkpad | zbenjamin: I wouldn't mind working on a different project for the next bit. Anything you need help with? | 06:59 |
akiva-thinkpad | you* as in the team working on the touch in general. | 06:59 |
zbenjamin | akiva-thinkpad: oh we have loads of work :D, just check out the bugtracker on a project you are interested in and start hacking. | 07:00 |
akiva-thinkpad | heh sure~ | 07:00 |
akiva-thinkpad | zbenjamin: anyways before I depart; I pushed another update that fixed a segfault, and resized the window to fit the background. Before I hop to a different project, what do you need done on the autopilot plugin? I just want to make sure I leave it in a nice state. | 07:02 |
zbenjamin | akiva-thinkpad: did you create a new MR for that? | 07:02 |
akiva-thinkpad | zbenjamin: I think so. | 07:02 |
akiva-thinkpad | zbenjamin: yes, sent to the ubuntu sdk team | 07:03 |
zbenjamin | akiva-thinkpad: cool, i'll check it out, you can always work on some of your steps you listed that you wanted to fix. I did not have time yet to figure out more | 07:03 |
akiva-thinkpad | okay sure; as long as nothing stands out. | 07:03 |
akiva-thinkpad | zbenjamin: Am I vetted enough to be able to be on the sdk team yet? | 07:04 |
zbenjamin | akiva-thinkpad: that is a question for bzoltan_, not sure if you can be without actually working for us | 07:04 |
akiva-thinkpad | zbenjamin: fair enough! | 07:05 |
zbenjamin | akiva-thinkpad: but even if you are not, its the same work flow, branch , hack hack hack, make merge request, get it approved. After that it makes no difference on who merges ;) | 07:06 |
zbenjamin | akiva-thinkpad: what you can always do is to write tests for your stuff | 07:06 |
akiva-thinkpad | zbenjamin: well the main thing is that I can't update the synopsis on the project page, which is now out of date due to new merges. | 07:07 |
zbenjamin | tell me what to change | 07:07 |
akiva-thinkpad | and I guess one final thing; as a developer, is there anything you think I need to work on? I really like the feedback. | 07:07 |
akiva-thinkpad | sec | 07:07 |
akiva-thinkpad | × Update of test suite on request (Low Priority) -- you can do that now. | 07:08 |
akiva-thinkpad | and you should remove × Get a Device, because I don't actually have one. (Medium Priority) -- as that refers to me | 07:08 |
akiva-thinkpad | I still don't have one :( | 07:08 |
zbenjamin | :( | 07:09 |
zbenjamin | akiva-thinkpad: well, as a developer if you want to work with C++, boost your knowledge about pointers. That was the main problem because it had lots of leaking stuff in the beginning | 07:10 |
zbenjamin | akiva-thinkpad: did you do java before? | 07:10 |
akiva-thinkpad | zbenjamin: yeah I was under the impression that a lot of the pointers auto-deleted. I also need to make a better habit of checking for null pointers | 07:11 |
akiva-thinkpad | thanks | 07:11 |
zbenjamin | akiva-thinkpad: so, read about QObject parentships | 07:11 |
akiva-thinkpad | zbenjamin: Will do! | 07:11 |
zbenjamin | akiva-thinkpad: they are deleted automatically if their parent is deleted. But you need to specify the parent of course ;) | 07:12 |
zbenjamin | akiva-thinkpad: so generally if you "new" something take extra care that either YOU delete it, or something takes ownership of the pointer | 07:12 |
akiva-thinkpad | yeah, and there is that attribute you can also set, that I didn't know about that will delete if you close the window. | 07:12 |
zbenjamin | akiva-thinkpad: yes for Widgets | 07:13 |
akiva-thinkpad | Good stuff. | 07:13 |
zbenjamin | then reading the Qt code if you have a problem with it teaches you a good deal of how to write code | 07:13 |
akiva-thinkpad | Anyways I won't keep you any longer. Oh you should also remove this: ✓ HUD/menu integration (left alt in unity7) | 07:14 |
zbenjamin | thats how i started years ago ;), but i did know how to write C and C++ | 07:14 |
zbenjamin | akiva-thinkpad: done , thanks for putting your time into that | 07:15 |
akiva-thinkpad | all great stuff. Thanks so much for the guidance and reviewing my code. I learned a ton | 07:15 |
zbenjamin | yw | 07:15 |
akiva-thinkpad | np! | 07:15 |
zbenjamin | so when you are interested in something different don't be afraid to ping the code maintainer, or ask balloons he can point you to the right people | 07:15 |
akiva-thinkpad | Cool. Later on I may ask you and some of the other devs I have worked with to sit as a reference on a job application. Is that alright presuming it isn't burdensome? | 07:17 |
akiva-thinkpad | It helps if I can show that sort of experience to other potential employers. | 07:19 |
zbenjamin | akiva-thinkpad: sure | 07:21 |
akiva-thinkpad | Great stuff! Thanks; it is extremely appreciated. | 07:21 |
akiva-thinkpad | Okay I gtg. | 07:23 |
akiva-thinkpad | cheers | 07:23 |
mivoligo | mzanetti: hi, what was the option to run MvsM in dev mode? | 07:45 |
=== chihchun_afk is now known as chihchun | ||
=== john-mca` is now known as john-mcaleely | ||
=== davidcalle_ is now known as davidcale | ||
=== davidcale is now known as davidcalle_ | ||
=== davidcalle_ is now known as davidcalle | ||
=== faenil is now known as faenil_ | ||
=== faenil_ is now known as faenil | ||
mcphail | Does anyone know whether qmlscene si going to be the long-term solution to launching qml apps? It is not a production-ready piece of software (and explicitly admits that), and is very slow. | 09:04 |
mcphail | "The qmlscene utility is meant to be used for testing your QML applications, and not as a launcher in a production environment." from http://doc.qt.io/qt-5/qtquick-qmlscene.html | 09:06 |
ogra_ | i heard it will be dropped soon in favour of an actual launcher | 09:08 |
ogra_ | me thinks he remembers mzanetti saying something like that | 09:09 |
mcphail | ogra_: that's great news. Think it will bring a huge leap forward in the usability of the phone | 09:09 |
ogra_ | if qmlscene gets renamed ? | 09:09 |
mcphail | ogra_: if qmlscence gets nuked from orbit | 09:10 |
ogra_ | i dont see what would change in the usability due to that :) | 09:10 |
ogra_ | it will just get a new name and likely function a little bit different | 09:10 |
mcphail | ogra_: that would be a disaster. Loading times for simple apps are too long | 09:10 |
mcphail | ogra_: apps without qmlscene are faster | 09:11 |
mzanetti | mcphail, it's really not the qmlscene binary that's slow, but the creation of the QQuickView and then the building of the QML object tree | 09:16 |
mcphail | mzanetti: is there work going on to profile that? | 09:16 |
mzanetti | not that I'm aware of | 09:17 |
mcphail | mzanetti: reason I'm asking is because some of the MX4 users have expressed disappointment that app loading is still slow despite much better hardware | 09:17 |
mcphail | mzanetti: so it seems crazy to rely on a launcher which admits it is not fit for purpose | 09:18 |
mcphail | personally, I cringe when someone asks to see my phone and it takes 5 seconds for the dialler app to open | 09:20 |
mzanetti | mcphail, fwiw, the dialer app doesn't use qmlscene | 09:21 |
mcphail | mzanetti: aah - what is the delay there then? | 09:21 |
mzanetti | mcphail, it's really not the qmlscene binary that's slow, but the creation of the QQuickView and then the building of the QML object tree | 09:22 |
mcphail | ha! | 09:22 |
mcphail | mzanetti: so, is the dialler a "normal" Qt app with a qml frontend? | 09:23 |
mzanetti | yes | 09:23 |
mcphail | mzanetti: so my frustration should be expressed upstream to the Qt people? | 09:23 |
mzanetti | maybe | 09:23 |
mzanetti | not sure how much that helps... but yeah, it's not ubuntu that's adding the slowness | 09:23 |
mcphail | That is enlightening. Thanks for correcting my misperceptions | 09:24 |
mzanetti | it's hard... I mean... maybe the dialer app could be optimized a bit still | 09:24 |
mzanetti | then in theory one could preload a QQuickView and hand it over (like MeeGo did), but our security team thinks that's too risky | 09:25 |
mzanetti | then there are ongoing threats on the qt devel mailing list about speeding up creation of some things | 09:25 |
mcphail | mzanetti: that was going to be my next question :( | 09:25 |
DanChapman | is the qml compiler still tied to qt version? That would help a *bit* if that's no longer a restriction | 09:25 |
* DanChapman googles | 09:26 | |
mzanetti | we have the qml compiler in use afaik, well a self made one afaict | 09:26 |
mzanetti | that did improve things by a second or so | 09:26 |
ogra_ | mcphail, if an app starts slow on the MX but fast on the other phones, it is surely not qmlscene at fault ;) | 09:27 |
mcphail | ogra_: app don't start quickly on the bq | 09:27 |
ogra_ | the hardware and driver side of the MX4 simply still needs a lot of love | 09:27 |
mzanetti | I'd be happy if I'd be able to flash that thing :D | 09:28 |
greyback | mzanetti: we use qml caching - not the same as compiling. But caching does improve startup perf | 09:28 |
ogra_ | (if you have the ability to use both phones side by side you would actually agree ;) ) | 09:28 |
mcphail | ogra_: :) | 09:28 |
ogra_ | mzanetti, OTA works fine, you only need to get over the hump once ;) | 09:28 |
mzanetti | ogra_, what do you think I'm doing here? | 09:29 |
ogra_ | lol | 09:29 |
ogra_ | get a USB 1.1 HUB ;) | 09:29 |
ogra_ | (definitely helps if you slow down the USB speed) | 09:29 |
mzanetti | hmm.. interesting. maybe I could flash it from an RPI | 09:30 |
ogra_ | lol, i never tried that | 09:30 |
mcphail | I've got a Baldur's Gate install which is a 137MB package, opening 2GB of data from the SD card. It opens instantly when I click the icon. But a QML app I've written of a few dozen lines of code takes 4--5 seconds to open. Something is badly wrong there | 09:30 |
ogra_ | not sure we even have a ubuntu-device-flash binary for armhf | 09:30 |
mzanetti | mcphail, and you think thsoe 2GB of data are loaded at startup? | 09:31 |
mcphail | mzanetti: no, but most of the 137MB is | 09:31 |
mzanetti | not sure about that | 09:31 |
mcphail | mzanetti: I compiled the package... | 09:31 |
popey | The other C++ apps also start instantly. | 09:31 |
mzanetti | well... again... it's loading the QQuickView and building the object tree. It's a bit like opening a browser and loading a (local) website | 09:32 |
mivoligo | mzanetti: what was the command to run MvsM in developer mode? I have to fix that level 23 ;) It's harder on easy than on hard | 09:40 |
mzanetti | -d | 09:40 |
mzanetti | lol | 09:40 |
mzanetti | how did you manage to do that? | 09:40 |
mivoligo | I think is because the enemies on hard synchronize better to tower shots on that level | 09:41 |
mivoligo | mzanetti: what should be before "-d"? :P I really don't remember | 09:44 |
mzanetti | qmlscene -I path/to/builddir/backend/ path/to/machines-vs-machines.qml -d | 09:45 |
mivoligo | thanks! | 09:45 |
mivoligo | normally I just start from SDK | 09:46 |
mzanetti | mivoligo, you can add the -d to the sdk | 09:46 |
mzanetti | in project -> run settings -> additional arguments | 09:46 |
mivoligo | mzanetti: great, that's useful :) | 09:46 |
=== chihchun is now known as chihchun_afk | ||
=== alex-abreu|off is now known as alex-abreu | ||
=== _salem is now known as salem_ | ||
=== javiercrowsoft1 is now known as javiercrowsoft | ||
mzanetti | popey, rpadovani, joining? | 15:33 |
popey | omw | 15:33 |
nik90 | bzoltan_: ping | 20:34 |
nik90 | bzoltan_: (leaving messages for you to read when you come online) | 20:34 |
nik90 | bzoltan_: Is there anything holding up https://code.launchpad.net/~zsombi/ubuntu-ui-toolkit/alarmmodel-segfault/+merge/262074 ? | 20:35 |
nik90 | bzoltan_: zsombi_ told me we could get it into OTA-5. Is that still the plan? | 20:35 |
=== pat_ is now known as Guest94669 | ||
=== athairus is now known as athairuzzz |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!