| 15 Nov 2024 |
 atemu12 | Oh, that I don't know | 23:50:11 |
| atemu12 | They will likely supply Chrome but I don't know if they supply webview-chrome | 23:50:27 |
| atemu12 | There are some extra parts you need in order to be a webview IIRC | 23:50:56 |
| atemu12 | You can also install a webview using a Magisk module which is what I resort to currently | 23:51:11 |
| atemu12 | I'd really rather have a presigned webview provided by robotnix and let that update from upstream via F-droid | 23:51:49 |
| 16 Nov 2024 |
 oak 🏳️🌈♥️ | I needed to manually set lower jobcount for that mk_repo.py script because apparently I started getting 429 from the services because 32 threads were bombing the git providers | 00:05:49 |
| atemu12 | Google will time you out if you're run it a few times | 00:06:20 |
| atemu12 | I don't think it matters how many threads | 00:06:29 |
| atemu12 | I'd recommend you to run the repo and device dir stuff separately | 00:06:48 |
| oak 🏳️🌈♥️ | I just added --jobs 3 there into the update.sh script and now it seems to work | 00:07:15 |
| atemu12 | Huh okay | 00:07:21 |
| atemu12 | Good do know | 00:07:37 |
| oak 🏳️🌈♥️ | I think it could be a good idea to put some maximum in that default if it's based in CPU core count | 00:13:28 |
| atemu12 | I think repo does that by default | 00:13:48 |
| oak 🏳️🌈♥️ | https://github.com/nix-community/robotnix/blob/master/scripts/mk_repo_file.py#L251 | 00:15:32 |
| oak 🏳️🌈♥️ | It's not like the git cloning problem is that much CPU bound anyway, the most speedup comes from if some jobs wait for remote doing something while others can transfer during that time | 00:16:58 |
| atemu12 | You'd be surprised how CPU-bound it can be | 00:18:27 |
| atemu12 | Git needs to undo the delta compression for a whole bunch of BLOB repos and kernels | 00:19:01 |
| atemu12 | When I run the updater on my little home 4C server, it readily maxes out the CPU | 00:19:58 |
| oak 🏳️🌈♥️ | Now it's time to test, lineage-19.1 build | 00:33:35 |
| oak 🏳️🌈♥️ | * Now it's time to test, lineage-19.1 build (without webview) | 00:33:46 |
| oak 🏳️🌈♥️ | I wonder if I should create issues for all the things I encounter | 00:35:47 |
| atemu12 | Depends on what they are | 00:36:16 |
| atemu12 | There's a matrix for stuff that's expected to work or not | 00:36:27 |
| oak 🏳️🌈♥️ | It wasn't actually the webview that was failing, I think it was the chromium apk package itself | 00:37:00 |
| atemu12 | Yeah that's not unexpected | 00:37:17 |
| atemu12 | I created an issue on yeeting chromium earlier | 00:37:31 |
| oak 🏳️🌈♥️ | Ah yes this one I did encounter before:
error: hash mismatch in fixed-output derivation '/nix/store/hcz1f4ld47syakxfkl54p6s3g3jr972i-android_external_chromium-webview_prebuilt_arm-3a32a73.drv':
specified: sha256-XEzHw0c/xPM3X8VDptJNZln2aOlxhkRXp+EDp1eM4NI=
got: sha256-C9Syc7A5OyL4iGV1n9qNvwsZ//A5t3E5PDOR0wd5Mu8=
For all these webview prebuilts the hashes are wrong, these should get updated by the update.sh?
| 00:39:13 |
| atemu12 | Nope, webview stuff is entirely separate | 00:39:39 |
| atemu12 | update.sh is for lineageos onlu | 00:39:45 |
