Automated builds (V)Machine

Since building requests are quite numerous (and, of course, impossible to satisfy all), it would be cool to set up a VM (such as a jupiter notebook) that can - if correctly feeded - perform sources’ compile/assemble to obtain an unofficial custom build.

Some interesting links that may inspire:

Hope that inspires !

3 Likes

OK, we’ll use this 3ad as a notepad to pin interesting resources (maybe) useful for building/porting automation:

TWRP tree

LOS tree

Blobs

1 Like

Another VERY interesting find:

ROM-builders organization is created to support and help collaborate developers working on custom Android ROMs that don’t have access to powerful enough computers to build AOSP locally in a timely manner.

This organization provides powerful servers (sponsored by Cirrus Labs) configured and optimized for building AOSP fast

https://github.com/ROM-builders/how-to-add-your-rom

.

2 Likes

…some interesting links that may be a part of the “build automation”:

This post in XDA Forums has details of how to get vendor blobs for Sony Xperia devices.
https://forum.xda-developers.com/t/guide-extract-vendor-from-stock-firmware.4212587/

The sensible way to do this is to extract the files once, and store them somewhere they can be accessed by the extract-file.sh script in your build. The extracted files wont be changing so there’s no point repeating the process every time you build.

Even easier, once you have run the extract-file.sh script and the extracted files and generated makefiles are in the vendor\<manufacturer>/<device> directory, upload them to a git repo, where they can be accessed during repo sync

You’re right, but seems that some manufacturer obtained the bust of repos that (illegally ?) share proprietary extracted files and AFAIK vendor files may also change with stock rom upgrade.

Perhaps the best approach would be to make the user choose which link to provide: either the one to the already extracted files or the stock rom link (to extract vendors from).

(Appetite comes by eating) In the second case it could also be interesting if the extraction script would be able to publish processing results on a repo of the user who executes it.

This topic was automatically closed after 30 days. New replies are no longer allowed.