Code review comment for lp:~parthm/bzr/262450

> On 18 February 2010 00:18, Parth Malwankar <email address hidden> wrote:
> > Based on Martins comments:
> >
> >>So probably the permissions stuff just can't work like this on Windows, in
> which case we should probably:
> >> 1- make that test requireFeature something unixy; probably it would be good
> to declare a new Feature for unix permissions and just say it exists if
> os.platform=posix
> >
> > Added PosixOSFeature and the permissions test case now requireFeatures this.
>
> I think, to be a bit pedantic, the actual check is for unix
> permissions; this could conceivably pass under cygwin and fail on unix
> with a vfat filesystem.
>

Added feature PosixPermissionsFeature which specifically checks for file
permissions support (by create-change perms-verify perms sequence) if
os.name is posix.
Tests are passing on Windows (a borrowed system) and Linux. Windows skips the
test as expected.
Haven't been able to verify on Cygwin or OSX and I don't have those systems.

Removed PosixOSFeature added earlier.

--
Parth