Hi Drew,

Some website funkiness blocked me from logging in and posting this follow up 
review of physical.py so I mailed it to you privately as you know. I'm just 
re-posting my review comments here to keep things together.

Cheers,
Niall

physical.py
-----------
130:
            try:
                partition.size = Size(size.get("val"))
                partition.start_sector = int(size.get("start_sector", 0))
            except:
                 # catch any failure
                 raise ParsingError("Partition size element has invalid " + \
                                   "'start_sector' attribute")
It seems that either of the two expressions could raise a ParsingError, yet
the error message assumes it's always the second one (start_sector).

281:
    def remaining_space(self):
        """ remaining_space() - class property to return a Size object of the
        remaining overall space available on the Partition
        """
Nit: Is this not an object (instance) property rather than a class properry?
299: Same as 281
308: Same as 281
317: Same as 281
1081: Same as 281


299: def is_logical(self):

Suggestion - not something I'll get hung up on :-)
How about making this just return the logical inverse of is_primary?
eg. if self.is_primary:
        return False

308: def is_extended():
Similar to above, how about:
    if self.is_primary() and \ 
        self.part_type in  Partition.EXTENDED_ID_LIST:
        
667:
        if 5 <= index <= 36:
            bootid = Partition.INACTIVE

Why are you not comparing against the const "const.FD_NUMPART" as used in
Partition.is_primary() eg.
        if const.FD_NUMPART < index <= 36
Where does 36 come from? Might be useful to add a short comment.

901:                # XXX does this command really work?
 Does it? :)

942, 947:
Nit:  number_parts = number_parts + 1 -> number_parts += 1
-- 
This message posted from opensolaris.org
_______________________________________________
caiman-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/caiman-discuss

Reply via email to