Secret sexdate noreply
Because I am a bad maintainer things got messy this past merge window with respect to code review, merge, and design.
Since we are potentially entering a new era with many more contributors I'd like to try to structure things a bit more formally to prevent this from happening in the future.
I was hoping to get this thing finished by tonight, but its taking longer than I thought.
If you are interested in looking at a snapshot you can grab the PDF at browse the HTML version at I also did a doxygen pass on the v9fs relevant directories -- while our comments are doxygen formatted, it does a nice job of graphing the structure and cross-referencing the functions.
This is an unfinished work, but comments and contributions are always welcome. O curso a distância da FATESI têm ajudado Pastores(as), Obreiros(as) e amantes da Teologia Bíblica.
The "right" thing may be to go looking at fixing aio to be more useful inside the kernel and fixing socket support -- but all signs point to that being a deeper rat hole than I want to get into at this point and if such a solution is eventually available it should be fairly easy to switch to it.
One of the things I was thinking of doing was pulling out and generalizing the async I/O routines in mux.c/trans_fd.c -- it seems to me they might be more generally useful to the kernel since a kernel async-io (particularly one that works with iovecs) seems to be illusive.
This feels a bit like backtracking since we just combined them -- but I think the mistake before was that we combined the mux for the transactions/tags with the multiplexer for the sockets.
Proposals for changes to this organization or API must be submitted through the mailing list for a general review period of two weeks before any patches relating to the organization or API will be considered (although code supporting proposals are actively encouraged).
The template for API specifications should follow man-page format in much the same way the Plan 9 and Inferno system software is documented.