Hi,

This is my first report of what we've been doing regarding the fdisk redesign project.

1) What's been done in previous weeks, before SoC development officially started:

- Consolidate and agree on internal fdisk API and data structures with Petr and mainstream util-linux community.

- Isolate DOS specific label logic. This is important since DOS partition table context was deeply embedded in the heart of fdisk logic, making it impossible to implement a new API. Now DOS, like all other supported labels, has its own file and functions.

- Fixed two nasty bugs that are currently present in most distros. http://git.kernel.org/?p=utils/util-linux/util-linux.git%3Ba=commit%3Bh=50f6... http://git.kernel.org/?p=utils/util-linux/util-linux.git%3Ba=commit%3Bh=44d2...

- Generic code cleanups.

2) What's been done during week 1:

- Implemented and merged in mainstream the initial functions and data structures of the new internal API. http://git.kernel.org/?p=utils/util-linux/util-linux.git%3Ba=commit%3Bh=823f...

- Implement dynamic debug support. By reading environment variables ($FDISK_DEBUG) developers and users can quickly be informed of important events during that occur during fdisk's entire execution. http://git.kernel.org/?p=utils/util-linux/util-linux.git%3Ba=commit%3Bh=f7b1...

- Generic code cleanups.

3) What's immediately next? - Add device topology and geometry to the API. We have decided to go strictly with libblkid for the former, therefore fdisk will probably depend 100% on it. This enables us not to reinvent the wheel, specially since this blkid is also part of util-linux.

-- All in all I believe we have had a good, solid start. It really helps that Petr, I and util-linux maintainer are quite familiar with the code, thus making quick and important progress. Petr's willingness and knowledge really make a difference and couldn't be happier working with him.

Thanks, Davidlohr