It is necessary that Parrot's filesystem interface make user-visible the
_event_ of measuring the attributes of a path _or_ an already open
filesystem object (e.g. calling stat() or fstat()). It must also represent
the bundle of measurements returned as some kind of PMC. (handwave handwave)
I agree with that.
Perl 5 does this (fairly crudely) by returning a list of values from stat()
or lstat() which you can examine at leisure, and by special-casing the "_"
filehandle to mean "results of last stat".
We have stat() and lstat() working as well as perl 5 stat and lstat.
The above-quoted example is only plausible if it's shorthand for, e.g.:
$P1 = $P0.stat # or lstat
$P1.is_dir()
$P1.is_file()
Looking to this code, $P0.stat should return a Stat PMC object, so we
can call is_dir and is_file.
Other option is to specify that 'is_dir' and 'is_file' stat the file
everytime they are called.
For me, both options are reasonable. Chip's proposal have the advantage
of caching the info. Using is_dir and is_file to stat everytime, have
the advantage of not having a specific Stat PMC object.
Other option, is having one is_dir() and one cached_is_dir().
--
Alberto Simões - Departamento de Informática - Universidade do Minho
Campus de Gualtar - 4710-057 Braga - Portugal
