| dcaliste | chriadam: hello. It seems that my device survived playing with eglplatform_wayland.so ;) I've opened an issue in libhybris github about the crash I'm observing. | 08:08 |
|---|---|---|
| dcaliste | If you didn't have time to look at, it's mainly about a routine that is called after window deletion, but still using a pointer to this window. | 08:09 |
| chriadam | dcaliste: I saw that, thanks for following up. Andrew noted that lipstick-security-ui and voicecall-ui are (as far as he knows) the only apps which close all of their windows without subsequently closing the process | 08:09 |
| dcaliste | I can easily enough protect against the window access if NULL, but I'm wondering if doing this I'm just curing the effect but not the root cause. | 08:10 |
| chriadam | so it's possible that somehow the pointer gets initialised in such a way to make future runs (of the window close handling code) take a different codepath | 08:10 |
| chriadam | dcaliste: indeed, it does seem better to find out _why_ | 08:10 |
| chriadam | (although in general, guarding an unsafe pointer access sounds good also - but I guess in this case it's not supposed to be unsafe, suggesting an earlier issue / logic error in the code) | 08:11 |
| dcaliste | Yes, but it's implying to better understand the Wayland protocol. Do you know who I can invite to look at the issue ? | 08:11 |
| dcaliste | I can still continue to play with this wayland_window.cpp code though and try to figure out myself the code path taken by QtWayland. | 08:12 |
| chriadam | dcaliste: internal to the company, only Andrew den Exter. external? Robin Burchell (w00t), Giulio Camuffo (giucam), Gunnar Sletta (sletta). | 08:12 |
| chriadam | but given how old the qtwayland is that we're using, well, this problem might not even present if we used a newer version | 08:13 |
| chriadam | so unlikely externals would be interested in investigating, if it is somehow dependent on the qtwayland | 08:13 |
| chriadam | most likely the advice would be "please update, your qtwayland is ancient" ;-) | 08:14 |
| dcaliste | Yes, it's not clear if the code path is a wrong call from QtWayland or from the internal logic of handling window swaping in eglplatform_wayland from libhybris itself. | 08:14 |
| chriadam | in libhybris, Carsten Munk (Stskeeps) may be interested to help (again, external now), or frajo/mal/abranson perhaps | 08:15 |
| dcaliste | About QtWayland, I'm waiting for the discussion on January in community meeting to see if we can plan something together to sort out what was introduced in the fork and can be upstreamed and what is the work to rebase Jolla QtWayland on upstream at version 5.6.3. | 08:16 |
| chriadam | dcaliste: upstream changed a lot, and pretty sure Qt isn't accepting patches for 5.6.x any more | 08:16 |
| dcaliste | Well, we can keep them ourself in mer like for QMF. | 08:17 |
| chriadam | true | 08:17 |
| dcaliste | The idea would be to set up QtWayland in Mer from upstream @ 5.6.3 and add all necessary patches to match the current state. | 08:17 |
| chriadam | that would make sense, I agree | 08:18 |
| dcaliste | Anyway, thanks for the names. It was laso mainly to keep you informed about my findings. You may add a link in the internal bug to this Github issue https://github.com/libhybris/libhybris/issues/402 if some people internally have time to investigate and would like to comment. | 08:19 |
| chriadam | I think early next year (from Jan 2nd onward) I will probably be investigating this to some extent | 08:22 |
| chriadam | IIRC Joona got it added to the approved maintenance task list | 08:22 |
| chriadam | it will be a learning experience for me, as I don't know the details of that part of the stack at all | 08:22 |
| chriadam | probably you'll find the culprit before I do :-D | 08:22 |
| dcaliste | Oh, good news ;) So I'll publish all what I find on the Github issue. | 08:28 |
| chriadam | thanks! | 08:30 |
| chriadam | I have to head home - have a good night, and if I don't see you tomorrow, have a great christmas and new year! | 08:31 |
| chriadam | I will approve the qmf CRs tomorrow | 08:31 |
| *** Renault_ is now known as Renault | 14:18 | |
| r0kk3rz | kimmoli: is this so you dont have to shovel your driveway? do finns even do that in winter? | 17:49 |
| kimmoli | r0kk3rz: yep | 17:53 |
| r0kk3rz | i admire your laziness | 17:54 |
| kimmoli | https://www.youtube.com/watch?v=-ESz2eDB0-0 | 17:55 |
| kimmoli | when the driveway is like 60m ... | 17:56 |
| caitlin | theres a lot of chatter on that 'plough' :/ | 18:14 |
| tsglove | jesus... I am amazed at having so much snow... that a mini-digger is used! | 18:27 |
| tsglove | I would dig that for at least one month of my life. Life is short. Try something new! | 18:27 |
| kimmoli | well, then there is this guy https://www.khl.com/news/finnish-service-engineer-sets-mini-excavator-world-record/56986.article | 18:32 |
| tsglove | ^ nice!! Long drive! | 18:35 |
| tsglove | n00b question: is Sailfish OS based in a linux distro? | 18:36 |
| r0kk3rz | not really no | 18:38 |
| tsglove | r0kk3rz, thanks! It's that I am looking to buy a Sony Xperia, and obviously want to install Sailfish | 18:38 |
| tsglove | was wondering the... the "where" it came from | 18:39 |
| r0kk3rz | out of the remnants of meego | 18:39 |
| r0kk3rz | but it is basically its own distro, and not based on any other as such | 18:40 |
| tbr | and moblin and maemo and … | 18:40 |
| tbr | there is an awesomely convoluted ancestry graph | 18:40 |
| r0kk3rz | yeah cue epic tale of woe | 18:40 |
| tbr | oyyy veeeey! | 18:40 |
| tsglove | thanks guys... cool | 18:41 |
| tsglove | So, out of curisity: Say I wanted to develop an app for Sailfish (after having bought one)... any pointers? | 18:44 |
| r0kk3rz | download the sdk | 18:45 |
| r0kk3rz | go nuts | 18:46 |
| r0kk3rz | typically apps are written in Qt C++/QML | 18:46 |
| tsglove | Heck yeah! Going to order that Sony Xperia now | 18:46 |
| r0kk3rz | but python/QML is commonish too | 18:46 |
| r0kk3rz | or really, whatever you want so long as it supports gles and wayland | 18:47 |
| *** Renault_ is now known as Renault | 18:57 | |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!