Hi, The following commits have been made in the flash-kernel repository (a) 8436937e3c7470cfac1c4a44c8c4c796c10712fe Don't scan for Android boot; add to machine-db (b) c5e08d6036237e306e6c83d177c2ada5453394ba Compute Android required size properly (c) b3a05a0d6e72aa7a160d0ac0864e63b1aec1c7a5 Remove android_boot_minsize variable too (d) 9a677eef0af04e73bcb6122b0031aa3249e03d9f Add and require Android-Boot-Config field (AC100)
As the contributor of the Android code, I do not agree with those changes and see them as counter-productive. I reworked the initial contribution multiple times to make things more flexible, easy and simple, and all those changes make things more complicated and more inflexible. Thus, I'd like for these changes to be reverted. The following lists the changes and my specific reasons for disagreeing with them: (a) BREAKS devices The reason why the code scans for Android boot partitions is that they are not fixed and can vary, even on devices with the same machine name. (b) UNNEEDED bloat We do not need to calculate the size of the boot partition. The abootimg tool we use to update the boot partition already checks for correct sizes. The removal of the size checks was done on purpose by me in the updates to the first patch, because the code is much cleaner and because abootimg can more reliably determine whether an action will be successful than we. (c) [Depends on (a) and/or (b)] (d) UNWANTED The change (d) makes the configuration file required. I do not want this at all. We do not need a configuration file to work, and I explicitely made the configuration file optional in my code due to this reason. Yes, the comment was misleading and could lead one to conclude that it's needed on the AC100, but it's purely optional, and I'd like to keep it that way as it greatly simplifies things for testing and updates and various workflows I have on the device. I also do not agree with the stylistic changes in commit c6a71832f6145d73476a427556ce3670154d8dff (longer code without benefits, a comment split in two lines and other lines getting longer than the whole comment on one line), but those are only stylistic changes, and do not break actively used functionality, and can be discussed later. -- Julian Andres Klode - Debian Developer, Ubuntu Member See http://wiki.debian.org/JulianAndresKlode and http://jak-linux.org/.
pgpfwmS29bG2t.pgp
Description: PGP signature
