| deathmist | could http://releases.sailfishos.org/sdk/installers/3.7.4/Sailfish_OS-Jolla_SDK-4.3.0.12-Platform_SDK_Chroot-i486.tar.bz2 get a .md5 like SDK 3.6.6 (SFOS 4.2.0)? | 01:39 |
|---|---|---|
| Wikiwide | Mister_Magister, piggz: Google Drive link doesn't work, could somebody share a working link, please? https://community.fxtec.com/topic/2467-community-build33016updated-18th-aug-sailfish-os-for-fxtec-pro1/ I do have Sailfish 3.2 somewhere on my laptop, but it would be more convenient to download 3.3 directly? If it is available... Otherwise, I would have to stumble around release updates, and that could brick my | 04:09 |
| T42_ | <xdfce> No one ,good luck | 04:26 |
| T42_ | <xdfce> https://sailfishos.org/content/uploads/2021/06/SailfishOS-HardwareAdaptationDevelopmentKit-4.1.0.0.pdf (re @BenKerry: Hiiii i have a Redmi...) | 04:26 |
| *** kimmoli_ is now known as kimmoli | 07:11 | |
| deathmist | mal: am I missing something or does one have to 'echo a | build_packages.sh --mw' to avoid the prompt when building all MW packages? | 10:08 |
| mal | deathmist: it seems it might actually always ask once | 10:16 |
| mal | maybe there should be some --quiet option to silently build all | 10:17 |
| deathmist | mal: or --yes / -y? I don't mind what it's called really, I can me the PR though | 10:17 |
| deathmist | s/me/make/ | 10:18 |
| mal | yeah, there are different options for the name, not sure which one is best | 10:19 |
| mal | I think both yes and quiet are used in different programs, not sure which one is more common, I somehow feel quiet would be more commonly used | 10:21 |
| deathmist | mal: https://github.com/mer-hybris/droid-hal-device/pull/318 :) | 10:41 |
| mal | deathmist: one possible issue, now that might be position dependent, I mean in here it sets that variable https://github.com/mer-hybris/droid-hal-device/blob/master/helpers/build_packages.sh#L89 so if the options are in wrong order --quiet does nothing? | 10:48 |
| mal | maybe you need to check somehow in that --mw if the variable was already set | 10:49 |
| deathmist | mal: pushed fix for that, it should work but I'll also test locally soon | 12:26 |
| Thaodan | IMHO if you just want to build all it shouldn't ask just build all or pass a specific option. | 13:45 |
| Thaodan | Quiet doesn't really fit since there's still output | 13:46 |
| deathmist | Thaodan: this is useful for scripting purposes where I'd like to run stuff potentially in the middle of the steps, and I definitely have to do that already to e.g. install droid-config after it's build so that droid-hal-version and droidmedia builds don't fail | 14:01 |
| deathmist | and I can change the option to --yes as I already proposed earlier | 14:02 |
| deathmist | I'm tired of having to maintain a droid-hal-device submodule fork just to make the build pass in one go... so I will not do that anymore | 14:02 |
| deathmist | Thaodan: also what should I print as the unclean tree error in https://github.com/mer-hybris/droid-hal-device/pull/288? I'd just like it to finally be considered to be merged | 14:13 |
| deathmist | also also what should be done about https://github.com/mer-hybris/android/pull/123? do I drop libhybris so it's still a manual step you have to find in FAQ and do before building HAL? | 14:27 |
| b100dian | voidanix[m]: for some reason, adding this logging/side effect changes kernel output for me https://github.com/sailfishos-on-tucana/ngfd/commit/d3936b22a9fb71b9ae898ccb44b28d66cd79c9f7 | 16:39 |
| voidanix[m] | b100dian: could you elaborate a bit on "changes kernel output"? | 18:11 |
| T42_ | <b100dian> I dont have effectId different than say, 5 or 2 on my xiaomi ffmemless driver | 18:32 |
| T42_ | <b100dian> Sounds like a compiler optimisation voidanix[m] | 18:33 |
| T42_ | <b100dian> Can you add that commit and test if it changes your output with cached effects? | 18:33 |
| voidanix[m] | will do | 18:41 |
| voidanix[m] | b100dian: vibration now magically works with CACHE_EFFECTS... wot | 18:54 |
| voidanix[m] | `value of data` was printed only a few times, now it only prints `data again` now | 18:58 |
| T42_ | <b100dian> Nice! Still have no explanation why the compiler optimizes that out without logging, but we're on to something here! | 19:02 |
| Thaodan | I would ask Jusa or mal to review your PR before any optimisations are done that change later after review comments | 19:14 |
| T42_ | <b100dian> There was a bug in that custom_data initialization is 'optimised out' by the compiler, and the 'fix' currently as revealed that was.. adding logging. So not the PR adding optimisations.. The compiler removing code. | 19:18 |
| Thaodan | I know but e.g. if you add comments about why that bug occurs the logging might be redundant. My point is that comments might add change requests. | 19:22 |
| T42_ | <b100dian> Ah, I see. I will track down the cause nonetheless, but this is a step forward, knowing that it is optimised code. Logging should not be the final solution. | 19:25 |
| T42_ | <HengYeDev> I am facing this problem with my port: lineage-17 base works without audio, lineage-18 works with audio as it accounts for a hardware revision. should i try to backport the audio fix or use hybris-18.1 | 19:41 |
| Thaodan | HengYeDev: depends on the fix, if the fix is small backport but if its biger wait for Android 11 being fully working | 20:48 |
Generated by irclog2html.py 2.17.1 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!