stan | would you be interested in SDL1.2 and SDL2 modified to use the OK+ mapped keys in maemo-leste? (e.g. 'ESC') key? | 07:53 |
---|---|---|
stan | either SDL gets modified to use the remapped keys, or every game needs a leste version modified for N900 and Droid4 | 07:54 |
Wizzup | marex: hi! that would definitely be appreciated | 09:15 |
Wizzup | (I might be out parts of today, and Ivo ( freemangordon ) is gone for 2-3 more days, though | 09:16 |
Wizzup | stan: if you have patches for the maemo keymappings, you can submit the patch | 09:16 |
stan | i see sdl's sym->unicode gets our mapped keys but sym->sym does not | 09:18 |
stan | i'll try to find some relevant discussion in sdl mailing lists or issues | 09:20 |
stan | other devices probably face similar issues | 09:20 |
stan | the 5-level brightness applet should map to 4-7 without dynamic brightness | 09:48 |
stan | or 3-7 rather | 09:49 |
stan | how are you using your droid4's at all right now? it's either too dark or full-on bright. | 10:38 |
lel | clort81 opened an issue: https://github.com/maemo-leste/bugtracker/issues/549 (Getting SDL apps to recognize mapped keys (particularly Esc)) | 10:51 |
stan | <Wizzup> parazyd has taken surf, which uses webkit-gtk3 << nicely minimal and pinch to zoom works, but Xorg sits at 88% cpu on idle page | 13:24 |
parazyd | stan: It's possible that is happening because of opengl | 14:32 |
stan | idk. i just see a lot of modern software redrawing when it doesn't need to. that's a common culprit of the modern 'cycles are free' attitude | 14:35 |
parazyd | Yup | 14:36 |
parazyd | All mainstream web browsers/engines are like this | 14:36 |
stan | but i doubt the commercial phone browsers do it | 14:39 |
bencoh | surprisingly enough, I found firefox to give better results once you disable prefetch and a few other similar options | 14:41 |
bencoh | (oh and, I use it with ublock origin as well) | 14:41 |
parazyd | Yes it seems faster | 14:42 |
bencoh | this might sounds like pure heresy for maemo, but maybe we should add some configurable per-app SIGSTOP facility for unfocused app in hildon-desktop | 14:44 |
bencoh | -s | 14:44 |
stan | someone wrote this and it is great | 14:44 |
parazyd | https://github.com/maemo-leste-extras/sigstoped | 14:44 |
stan | sigstoped | 14:44 |
bencoh | oh, really? | 14:44 |
bencoh | that's great | 14:45 |
bencoh | I just tried it, and it really seems awesome | 14:53 |
stan | it could be a useful addition to some desktop environments, with a little gui. few know about it, or even the ability to SIGSTOP | 14:55 |
bencoh | I'd even suggest installing it by default with maemo and let the user decide whether to enable it / populate the blacklist | 15:00 |
stan | i'm having to force backlight on in a permanent while loop in bash | 17:37 |
stan | this was not a good state to leave backlight control in | 17:37 |
bencoh | backlight control is a bit funky at times here as well, yeah | 17:47 |
bencoh | I think it depends on whether the sensor faces the light source or not | 17:48 |
stan | i should complain about my own poor results though. i'm reducing wasted cycles in a game that could run nicely on droid4 | 17:54 |
stan | but this trivial 2x scaled blit from one surface to another is not working | 17:54 |
stan | bencoh: do you have a way to disable the charging notifications? | 18:51 |
stan | i *RUN* the device on USB power | 18:51 |
uvos | sigstoped has a gui: qsigstoped | 19:05 |
stan | nice | 19:06 |
lel | clort81 opened a pull request: https://github.com/maemo-leste/status-area-applet-battery/pull/6 (Removed battery full and disconnect charger popup notifications.) | 19:46 |
parazyd | wat | 19:47 |
Danct12 | wat indeed | 19:49 |
Danct12 | i guess someone was really annoyed by the notifications | 19:49 |
uvos | lol | 20:20 |
Wizzup | it's probably as a workaround for the d4 charge/discharge notifications | 20:25 |
parazyd | Yeah | 20:26 |
uvos | this is silly, but we should maybe conisder an option to hide all notifications, like android's do-not-disturb | 20:26 |
stan | I'm testing it now. It works. | 20:29 |
stan | These notifications are redundant to the green charging light. | 20:30 |
stan | Blocking useful/wanted notifications in order to stop being interrupted by redundant ones does not seem optimal to me. | 20:31 |
uvos | leste is not droid4OS so they are usefull regardless of the light | 20:31 |
Danct12 | charging led doesn't seem to work on the n900 though, so that notification is very important | 20:32 |
uvos | charging led should work on n900 | 20:32 |
uvos | leds generally work | 20:32 |
parazyd | uvos: Is there a way to get firefox-esr in portrait mode? | 20:32 |
uvos | parazyd: yes | 20:32 |
Danct12 | well sometimes there's this flickering red led thing | 20:32 |
parazyd | Do tell :) | 20:32 |
Danct12 | maybe something's wrong with my n900 setup | 20:33 |
uvos | parazyd: really i thought you only wanted confirmation :P | 20:33 |
parazyd | oh | 20:33 |
Danct12 | (i checked, charging led works perfectly fine on fremantle) | 20:33 |
uvos | Danct12: is should work fine on leste too | 20:33 |
parazyd | It doesn't on Pinephone | 20:33 |
uvos | parazyd: you broke it :P | 20:33 |
parazyd | (portrait ff) | 20:33 |
uvos | oh | 20:33 |
uvos | right | 20:34 |
uvos | Danct12: it worked for me when i reimplemented the led for n900 | 20:34 |
uvos | so idk | 20:34 |
uvos | ill try to check at some point | 20:34 |
uvos | (my n900 is perpetully empty) | 20:34 |
Danct12 | uvos it looks like it *does* work as i see the red led flickers | 20:34 |
uvos | ok | 20:34 |
uvos | parazyd: /usr/share/hildon-desktop/transitions.ini | 20:34 |
uvos | forcerotation = 1 | 20:34 |
uvos | or add firefox to whitelist = | 20:35 |
uvos | id reccoment forcerotation = 1 it makes everything nicer | 20:35 |
parazyd | Thanks | 20:35 |
uvos | (except breaking tklock but i dont use that) | 20:35 |
uvos | also tklock mostly works, it just looks a bit wrong if you rotate with it open | 20:36 |
parazyd | Yeah | 20:37 |
uvos | since latest mce you can avoid tklock if you want btw | 20:38 |
Wizzup | uvos: maybe we can make the d4 report another state than charging, like full | 21:03 |
Wizzup | we do the same on the n900 | 21:03 |
uvos | it dose report full | 21:03 |
uvos | just that it dosent last | 21:03 |
parazyd | Yeah it enters a fast loop | 21:03 |
uvos | kernel wise | 21:03 |
uvos | i guess the n900 driver reports full untill its below 90% again or something like that | 21:04 |
uvos | or it dosent use the battery while on usb | 21:04 |
uvos | cpcap cant do that (or rather it can but we are concerned about running it in a register state android dosent) | 21:04 |
parazyd | What about charging thresholds? | 21:04 |
parazyd | Perhaps in userspace we could let it drain a bit before charging again | 21:05 |
Wizzup | that still means it will happen every now and then, not ideal | 21:05 |
parazyd | True | 21:05 |
Wizzup | we want it to still keep the battery charged but not report charging/discharging from kernel | 21:05 |
uvos | that would mean the kernel has to lie | 21:05 |
uvos | which it really should not | 21:05 |
Wizzup | ok | 21:06 |
Wizzup | sorry, back later, dinner, will read backlog | 21:06 |
uvos | we could make the battery status applet report on charging/not charging based on if usb power is connected | 21:06 |
uvos | i think this is how android works too | 21:06 |
uvos | and is also how charging-mode now works | 21:07 |
parazyd | This sounds like it should keep some state in userspace? | 21:11 |
parazyd | Like remembering if USB is disconnected or not? | 21:11 |
uvos | so charging mode just shows charging if power supply has a type of USB or AC eg /sys/class/power_supply/usb/type and /sys/class/power_supply/usb/online of this device is 1 | 21:13 |
uvos | this works in all cases | 21:13 |
uvos | but has the downside that charging-mode will show charging, 100% instead of "full" | 21:13 |
parazyd | Yeah | 21:14 |
parazyd | That's just aesthetics though | 21:14 |
sicelo | Danct12: you're getting red led with Leste? At least in Fremantle, red led means something is not right | 22:17 |
Danct12 | yep, red led blinking without breathe effect | 22:18 |
Danct12 | but nonetheless, it still says it's charging and it does charge to 100 just fine | 22:18 |
sicelo | Oh, you probably mean amber/orange? | 22:20 |
sicelo | :-) | 22:20 |
Danct12 | it's red, not amber | 22:24 |
sicelo | Weird | 22:25 |
sicelo | I guess I've lost touch with Leste :-p | 22:35 |
sicelo | You say it charges with amber in Fremantle? | 22:35 |
bencoh | normal charge in fremantle is blinking "yellow" | 22:37 |
bencoh | do you refer to that one by "amber", or to another color? | 22:37 |
sicelo | I'll say yes (else the ladies will say men are color blind) | 22:50 |
stan | dpkg-shlibdeps: error: no dependency information found for /usr/lib/libSDL-1.2.so.0 (used by debian/drnoksnes/usr/games/drnoksnes) | 23:15 |
stan | ldd debian/drnoksnes/usr/games/drnoksnes |grep libSDL | 23:16 |
stan | libSDL-1.2.so.0 => /usr/lib/arm-linux-gnueabihf/libSDL-1.2.so.0 (0xb6cf1000) | 23:16 |
stan | why does dpkg-shlibdeps think libSDL-1.2.so.0 should be in /usr/lib and not in the correct place | 23:18 |
stan | grep -r doesn't show any textstring pointing to the wrong place | 23:18 |
stan | ahh no it's in the wrong place | 23:20 |
stan | pk dpkg-buildpackage succeeds now | 23:24 |
stan | but it segfaults whereas the previous build did not | 23:56 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!