!UUYziobKGGxpovWyAN:nixos.org

Robotnix

240 Members
Build Android (AOSP) using Nix | https://github.com/danielfullmer/robotnix74 Servers

Load older messages


SenderMessageTime
2 Jan 2025
@atemu12:matrix.orgatemu12That's not going to be an issue16:45:10
@atemu12:matrix.orgatemu12Building meaningful checks is the harder problem16:45:47
@oak:universumi.fioak 🏳️‍🌈♥️I saw your PR about enabling LineageOS's prebuilt webview and it taking a lots of disk space17:02:07
@oak:universumi.fioak 🏳️‍🌈♥️Did you check is the LOS prebuilt webview a trichrome-variant of the Chromium build?17:02:32
@atemu12:matrix.orgatemu12No but I solved that issue17:03:31
@oak:universumi.fioak 🏳️‍🌈♥️I think Trichrome is a build which has 3 APKs, one for Chromium, one for Webview, and one for shared common parts both of those use17:03:34
@atemu12:matrix.orgatemu12The problem was caused by robotnix; more specifically what I declared it to do17:03:45
@atemu12:matrix.orgatemu12LOS doesn't need a chromium because it has this webview browser17:04:03
@atemu12:matrix.orgatemu12And honestly anyone who wants a chromium for browsing purposes should install one at runtime17:04:20
@oak:universumi.fioak 🏳️‍🌈♥️(In addition to Trichrome, you can just build either standalone Chromium or standalone Webview, but that then takes more disk space)17:05:48
@oak:universumi.fioak 🏳️‍🌈♥️* (In addition to Trichrome, you can just build either standalone Chromium or standalone Webview, or both separately as standalone, but that then takes more disk space)17:06:00
@atemu12:matrix.orgatemu12It claims to be a "webview" apk17:06:16
@oak:universumi.fioak 🏳️‍🌈♥️but yeah, about the CI and testing, I think proper testing would actually be to install the build into some device, and try to run some tests with that17:08:54
@atemu12:matrix.orgatemu12Emulator would be ideal17:09:25
@oak:universumi.fioak 🏳️‍🌈♥️but even checking that stuff builds successfully is usually a good start17:09:25
@atemu12:matrix.orgatemu12Yes although, at this point, if I can't verify that something works as intended or the author convinced me that they have done so sufficiently, I'm not merging it anyways17:10:47
@atemu12:matrix.orgatemu12CI would be most useful for catching stupid careless mistakes17:11:17
@atemu12:matrix.orgatemu12Or enforcing standards17:11:31
@oak:universumi.fioak 🏳️‍🌈♥️It's also useful for finding regressions, for example if you change something on LineageOS 21 builds, something might break on LOS 20 or AOSP build17:12:19
@oak:universumi.fioak 🏳️‍🌈♥️Currently the situation is a bit messy, we have documentation which claims all kinds of features, which don't currently work, but it's not stated anywhere which things work and which don't. And things that aren't actively tested on CI will get bit rot17:13:03
@oak:universumi.fioak 🏳️‍🌈♥️Actually there is Status list and I don't completely agree on stuff on subject to removal -column :D The OTA updater does work and I do use it17:17:23
@oak:universumi.fioak 🏳️‍🌈♥️It's not that useful though, because after running OTA update I still need to manually reinstall Magisk and MindTheGapps through Recovery17:18:03
@oak:universumi.fioak 🏳️‍🌈♥️I don't know could I already bake Magisk and MindTheGApps into the build17:18:26
@atemu12:matrix.orgatemu12Fair point17:19:20
@atemu12:matrix.orgatemu12Please send a PR :)17:19:30
@atemu12:matrix.orgatemu12I don't know about Google's spyware but Magisk: Yes17:20:08
@atemu12:matrix.orgatemu12https://github.com/nix-community/robotnix/pull/26617:20:23
@atemu12:matrix.orgatemu12https://github.com/nix-community/robotnix/issues/1417:20:33
@atemu12:matrix.orgatemu12(I think you'd know more about embedding gapps into AOSP than I do given you've worked in the industry :) )17:22:09
@oak:universumi.fioak 🏳️‍🌈♥️Yeah they did it in one project in the past but I don't have access to the project's Android tree anymore so I could check how it was done17:22:54

Show newer messages


Back to Room ListRoom Version: 6