FreeLists / openbeosstorage / [openbeosstorage] Re: Storage Kit Plan

[Keith Poole <keef@xxxxxxxxxxxxxxx>]

Some (vague) timescales would probably be helpful - also let us know 
what bits you want us to do next.

Also, after reading the Giampolo book, and disassembling the Be storage 
kit and kernel code, it would appear that the vfs api is used by the 
kernel, and there is another :-( api, similar to the BSD syscall 
interface.  So the order of calling is:

        Application --> storage_kit --> syscall (into kernel mode) --> 
vfs --> BFS (or other) --> device driver --> device.

Should we be looking into documenting the syscall interface??

            Keith

Tyler Dauwalder wrote:

>I'm going to try to get a plan of action together for 
>Michael and Rich in the next few days. I'm thinking 
>it'll look something like this (off the top of my head):
>
>
>Milestone #1 (planning):
>----------------------------------------------------------------------
>Roughly document R5 vfs layer
>Design set of C functions used to interface with vfs
>----------------------------------------------------------------------
>
>
>Milestone #2 (pre-alpha):
>----------------------------------------------------------------------
>Implement vfs interface over POSIX
>----------------------------------------------------------------------
>
>
>Milestone #3 (alpha):
>----------------------------------------------------------------------
>Implement some basic subset of the Storage Kit classes
>with our vfs interface, probably something like BEntry, 
>BEntryList, BDirectory, BFile, BNode, BNodeInfo,
>BPath, and BStatable, 
>----------------------------------------------------------------------
>
>
>Milestone #4 (alpha):
>----------------------------------------------------------------------
>Implement the rest of the Storage Kit classes
>----------------------------------------------------------------------
>
>
>Milestone #5 (beta):
>----------------------------------------------------------------------
>Reimplement C vfs interface with syscalls into the
>now ready kernel
>----------------------------------------------------------------------
>
>
>Milestone #6 (production)
>----------------------------------------------------------------------
>Polish it all up
>----------------------------------------------------------------------
>
>
>I'm going to look a little more closely at the classes and see
>if having only two milestones for the implementations makes
>sense, or if it's really going to be a bigger job than that.
>
>I also think continually testing and verifying whatever
>we're working on as we go would be a good idea, hence
>the lack of an explicit testing phase. Testing isn't much
>fun, but being as an integral part of the operating system
>as the Storage Kit is, I think it's really important that we 
>do a good job. 
>
>Now, my question for all of you is, does this sound
>reasonable? Is it too much, too little, or about enough
>work for any given milestone? (Not that it matters a whole 
>lot other than giving a consistent feel of progress and 
>accomplishment, but still... :-) Any other comments?
>
>-Tyler
>
>
>
>
>
>