Hello Zdenek,

thanks for your status report. I have reviewed your code and it is very promising. A few of my suggestions what to do next (except, of course, starting to write the text of the thesis):

* You should move the FUSE file systems to a dedicated directory (say
  uspace/srv/fs/fuse) in order to avoid confusion with the native FS
  drivers in the source tree.

* The code cleanup is indeed necessary. Except for deciding which TODOs
  and FIXMEs are you going to tackle eventually, you should also make
  sure that the original libfuse sources stay as much in-tack as
  possible (since you are not reimplementing them, you are reusing the
  vanilla libfuse), in order to be able to upgrade to a newer version
  as easily as possible.

  This does not necessarily mean that every feature of libfuse needs to
  be fully implemented. But the unimplemented features should be worked
  around by supplying dummy and empty POSIX required functions and
  macros to them, and preferably not by commenting out parts of libfuse.

* You should tackle the ticket #480 properly, not just by commenting
  out the locking. I know that this won't probably make you too happy,
  because it is an unforseen complication, but I believe that it is
  necessary.


M.D.

_______________________________________________
HelenOS-devel mailing list
[email protected]
http://lists.modry.cz/cgi-bin/listinfo/helenos-devel

Reply via email to