*** zbenjamin is now known as Guest91993 | 02:18 | |
*** zbenjamin_ is now known as zbenjamin | 02:18 | |
*** frinring_ is now known as frinring | 02:54 | |
*** nyov is now known as Guest91283 | 04:09 | |
*** jpich_ is now known as jpich | 09:36 | |
Tekk_ | Since I'm stuck with the platform sdk until the IDE finishes, is there any way to test graphical apps from the platform sdk aside from generating an rpm and slapping it on my phone? | 16:25 |
---|---|---|
ggabriel | build a 486 rpm and slap it in the emulator that comes with the sdk? | 16:28 |
Tekk_ | ggabriel: Can't install the sdk on modern distros though | 17:17 |
Tekk_ | Not without digging up whatever version of Fedora last shipped openssl1.0. Guess I might be able to try and dig through the .run file to extract the vm image. | 17:17 |
ggabriel | Tekk_: huh? really? I'm using Gentoo latest/stable and it works very well | 17:18 |
ggabriel | ...but you could extract the vm image, yes; should work on any vbox | 17:19 |
Tekk_ | gentoo has an openssl 1.0 package afaik | 17:21 |
Tekk_ | Fedora, Debian, and Ubuntu stopped packaging 1.0 years ago | 17:21 |
Tekk_ | You have to install packages from old releases for it to work | 17:21 |
ggabriel | I use 1.1.1g fwiw | 17:22 |
Tekk_ | I'm just scp'ing it over to my phone. | 18:09 |
Tekk_ | This experience isn't bad at all tbh c: | 18:10 |
gmc | Tekk_: all i had to do on ubuntu 20 was provide an older version of openssl, works fine after that | 18:31 |
gmc | and I believe that was only for the installer | 18:32 |
Tekk_ | Probably. The issue is in the installer's linkage afaik | 18:36 |
Tekk_ | The workflow with the platform sdk's been fine though | 18:36 |
Tekk_ | I don't like IDEs anyway, the only part I miss is the nice interaction with the vm c: | 18:37 |
attah | Tekk_: sfdk should be able to replace most of the SDK convenience with "just an editor" | 18:41 |
gmc | it depends.. i really love the jetbrains' ide's, they're so powerful | 18:41 |
attah | I should have put in the caveat "for building and deploying" | 18:43 |
pketo | Tekk_: fedora has compat-openssl10, does that help? | 18:43 |
gmc | gha.. so i define a 'signal taskUpdate(var task, string calendarUrl)' in one Item, then instantiate it in another, and get a 'Cannot assign to non-existent property "onCalendarUpdate"' on my 'onTaskUpdate: taskUpdate(task, calendarUrl)' | 18:43 |
* Tekk_ facepalms | 18:44 | |
Tekk_ | Yeah it technically does. I'm comfier working in emacs though :) | 18:44 |
Tekk_ | I might set up the IDE just to get the VM working all nice | 18:44 |
attah | Tekk_: That's an IDE... a elisp IDE ;) | 18:45 |
pketo | I don't know if there is easy way to get the emulator without the IDE | 18:47 |
attah | Hmmm? http://releases.sailfishos.org/sdk/emulators/ | 18:49 |
Tekk_ | The emulator integration though. | 18:50 |
Tekk_ | Though I'm guessing it's just some 9p deal like with qemu/kvm? | 18:50 |
gmc | hmm ok, lovely.. so when I move files around in my app, the file at the old location will stay around ... | 18:50 |
Tekk_ | That doesn't sound right? | 18:51 |
Tekk_ | Seems like rpm should take care of that for you | 18:51 |
gmc | that is.. when I have an older version with foo/bar.qml, then remove bar.qml too baz, i get baz/bar.qml, but the old version is still in foo/bar.qml.. | 18:51 |
Tekk_ | making good progress on my project btw. Once the hackathon is over I'll try and convince legal to let me release it :) | 18:51 |
gmc | Tekk_: yeah you would expect this to be handled somehow right? | 18:52 |
attah | gmc: i have seen that too... sometimes it even kept a broken app working for me, which i then went on to publish | 18:52 |
gmc | attah: so this basically means I cannot refactor my qml tree, and have to live forever with my old mistakes.... unless i can find a trick | 18:54 |
attah | yeah, one would have thought deploy as rpm would help... but i ended up just doing a few uninstalls | 18:55 |
gmc | attah: you think this is only in the emulator, not on the device? i do suspect that deploying throught the ide / sfdk doesn't actually do an update, but just force installs the package | 18:55 |
attah | gmc: definitely the same on real device, I haven't touched an emulator in years | 18:56 |
gmc | attah: ok, sorry, i meant 'deploy from ide or with sfdk' vs 'do an update (from openrepos or with rpm itself via ssh)' | 18:57 |
attah | aha.. could very well just be how sdk/sfdk does it | 18:58 |
attah | your theory seems reasonable | 18:59 |
gmc | i'm trying to validate it :) | 19:02 |
gmc | well, yes, when I `rpm -U` manually the old file *is* deleted.. | 19:04 |
attah | ...as expected.. but does the different upgrades behave, well, differently? | 19:05 |
gmc | sorry, no... it does *not* remove it either | 19:06 |
attah | strange | 19:06 |
gmc | yes, especially given that rpm is supposed to figure this out.. so now i'm suspecting my rpm from somehow being messed up | 19:07 |
Tekk_ | gmc: rpm doesn't know about files that aren't in its database though. Maybe one of the methods circumvents rpm? | 19:18 |
Tekk_ | and that's how you end up with your zombie files | 19:18 |
gmc | nah, i think what's happened is i installed the app with a version 0.7-1. then i moved the files around, and redeployed with the same version | 19:23 |
gmc | i just cleared the app from the emulator, then installed 0.7-1 (with the old tree structure), then checked out the commit with the new tree and installed that but with a bumped version (0.7-2) | 19:24 |
gmc | that does remove the old files | 19:24 |
Tekk_ | Ahhh | 19:24 |
attah | Seems it would be a good idea if at least the SDK could supply some switch to force reinstall behaviour even within the same version | 19:27 |
gmc | i suppose that would have saved me some time :) | 19:31 |
Tekk_ | Is there any API to bypass silent mode atm? | 19:43 |
Tekk_ | Well, silence and notification sounds at low/zero volume in general. | 19:44 |
attah | I assume you mean other than playing non-notification sounds? (: | 19:44 |
Tekk_ | In that case I'd have to bypass *media* volume being low/zero ;) | 19:44 |
Tekk_ | So either way | 19:44 |
Tekk_ | The notification docs mention a "critical" urgency but it says nothing more than "it's more likely to be displayed". Seems like there's no way to override notification sounds anyway so I suppose it'll be a media play regardless. | 19:47 |
Tekk_ | I guess that might just be mpris/dbus to set the volume to 100, then restore it once the sound is done playing? | 19:48 |
attah | Tekk_: i guess you read this: https://github.com/CODeRUS/android-notifications/blob/master/qml/pages/MainPage.qml | 19:54 |
Tekk_ | ....coderus' android notifications supported custom notification sounds? I *never* received a custom notification sound from an android app. | 19:55 |
Tekk_ | It was always the normal message sound | 19:55 |
attah | i dunno.. never used it, just looks promising | 19:55 |
attah | it also somehow seems to respect notificationEnabled - which leads me to believe it can be disrespected | 19:56 |
Tekk_ | Seems like it. I think I'd need to mess with volume for a low notification sound volume thing anywya, though. | 20:00 |
Tekk_ | Whole point is that a notification from the app has to be able to wake me up when I'm sleeping :) | 20:01 |
attah | Jolla 1 with Klaxon TOH? | 20:01 |
Tekk_ | XA2 Ultra unfortunately :p | 20:02 |
Tekk_ | Though I have pulled a nice alarm ogg. the US navy released the "we're all going to die" alarms recorded off of submarines. Just gotta play it at full volume regardless of volume settings. | 20:02 |
coderus | oh my hacky-dirty project :)) | 20:15 |
attah | the best kind of project :) | 20:20 |
Tekk_ | coderus: Might've been, but it was awfully helpful before sailfish supported android app notifications properly :) | 20:24 |
Tekk_ | Gross. Looks like pulse doesn't actually provide a dbus interface and mpris doesn't let you do system volume control. | 20:43 |
Tekk_ | Huh. I found a unicorn. | 20:59 |
Tekk_ | CodeReader is horribly out of date, broken, and abandoned in *openrepos*, not the other way around for once :) | 20:59 |
attah | Are you sure you are looking at the same incarnation of it? | 21:00 |
piggz | rinigus: so, i fixed the combobox issue by moving the assignment of the item to the combobox.oncompleted ... however, i dont like not knowing why is didnt work in the page.oncompleted, or on a timer | 21:01 |
Tekk_ | Different incarnations, hence the issue attah | 21:01 |
Tekk_ | the author on openrepos is a steffen_f, and that version no longer works. hasn't been touched since 2017 | 21:02 |
Tekk_ | Just to double check since it's a qt component and not a sailfish one: A timer ought to trigger even if the app isn't foregrounded, right? | 21:22 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!