| rinigus | piggz and mal: found the source of a failure at SB2 during qt webengine compilation (after perl workaround). described at https://forum.sailfishos.org/t/sb2-fails-to-mmap-8gb-in-aarch64/15159 . note that on device, I can use mksnapshot compiled by SB2 and generate required .S and .cc files. | 07:03 |
|---|---|---|
| rinigus | but let's see, if Jolla SB2 devs can help to fix the issue | 07:03 |
| ViGe | Sure, if you give us a few years :( | 07:04 |
| rinigus | Nooh, its a simple example, isn't it :) . | 07:04 |
| ViGe | I fear the issue is because the SDK is still 32 bit, so using more than 4GB of memory is pretty much no go | 07:04 |
| rinigus | That could be. hmm, maybe then I have to see what will happen if I just lower allocated RAM requirement | 07:05 |
| rinigus | any plans to move SDK to 64bit | 07:05 |
| rinigus | ? | 07:05 |
| ViGe | Yes, there are plans, but as usual there is no timeframe. | 07:06 |
| rinigus | ViGe: thanks for insight - I haven't thought that SDK is 32 bit. will have to think how to work around it and I am sure we will find the solution. | 07:07 |
| rinigus | in principle, I can generate codes on device and add them as source. then just insert into the build at the right moment. | 07:07 |
| ViGe | That's one option. Not a good one, but then again I can't really suggest anything better either :/ | 07:08 |
| rinigus | or reduce RAM allocation while generating these files. just will have to figure out how while keeping it working for device in full swing | 07:08 |
| rinigus | probably can compare if the device generated and files generated with smaller RAM requirement would be the same... | 07:09 |
| rinigus | will see. at least the problem is known now. :) | 07:10 |
| dracks[m] | <mal> "one way to check for memory..." <- Thanks, I was taking a look, I get it run, but the output was weird, the app was never shown in the device, and inside the valgrind, there was at some point an error of an unknown instruction executed, after that, I've got a lot of "warnings" about thinks complettly unrelated to my app, like warnings of the android or wayland... also some of them from Qml itself... When I worked long time ago a | 07:32 |
| dracks[m] | little bit with c++, I knew that you need to allways create a destructor for you classes, otherwise C++ was not enough smart to call the destructor of the internal variables, the problem is still there? Thanks! | 07:32 |
| piggz | rinigus gets +1 in debugging skills | 08:00 |
| rinigus | piggz: thanks! great wiki at https://github.com/sailfishos-chum/qt5/wiki/Getting-Started . please add link to README at https://github.com/sailfishos-chum/qt5 | 08:26 |
| ViGe | piggz, rinigus, you guys are moving so fast it's really hard to keep up ;) | 08:30 |
| rinigus | ViGe: we are just cutting corners ... | 08:34 |
| piggz | ViGe: a project born of frustration ;) | 08:48 |
| *** amccarthy is now known as Guest8636 | 10:27 | |
| *** amccarthy_ is now known as amccarthy | 10:27 | |
| mal | @dracks[m] it's normal to see warnings from android side in valgrind | 11:24 |
| mal | rinigus: I probably found the root cause for the local build issue of qt packages, not sure yet where it needs to be fixed | 15:18 |
| mal | it seems it might have been a false lead | 15:29 |
| piggz | mal: progress looking good, 2 packages to go https://github.com/sailfishos-chum/qt5/issues/1 | 15:47 |
| b100dian[m] | "only" an entire browser to go:) | 16:13 |
| piggz | b100dian[m]: and only for aarch64 ... he showed me the browser working in the emulator :) | 16:17 |
| rinigus | piggz: I unticked webengine - still in progress. | 16:53 |
| rinigus | getting closer, though | 16:53 |
| poetaster | rinigus, so wasm in the browser, eh ? (ducks and runs) | 16:54 |
| pherjung[m] | I'd like to automate some of the manual operations in the SIM toolkit. Is collecting logs using Ofono Logger a good way to get started? | 17:11 |
| pherjung[m] | thanks in advance : | 17:11 |
| pherjung[m] | * advance :) | 17:11 |
| b100dian[m] | poetaster: we have wasm in the browser at home | 20:19 |
| poetaster | b100dian[m], I have wasabi in my browser. it's not having the effect I hoped for. But, with ginger, it's palatable. | 20:30 |
| b100dian[m] | Really, it passes some of these checks, which is not a small feat https://wasm-feature-detect.surma.technology/ | 20:38 |
| b100dian[m] | But I suppose you have an use case that doesn't work | 20:39 |
| poetaster | nah, I was expecting mozilla to spawn GPT-3.5 by accident a year ago. The browser as operating system thing went off the rail long ago. So, it won't surprise me. | 21:29 |
| poetaster | Still, I should stop dumping browser view apps into the ecosystem :) except for the mod tracker. that thing is way. | 21:29 |
| b100dian[m] | You definitely don't even need protected mode for a mod tracker, I've seen the 90s. You could do it in the bootloader :) | 22:06 |
| b100dian[m] | The challenge is to have that wired speaker hmm | 22:07 |
| b100dian[m] | Speaking of browser view apps, any hydrogen user on the matrix side of this channel? | 22:08 |
| *** foobar is now known as Guest8682 | 22:48 | |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!