>>>>> "Martin" == Martin Gasbichler <gasbichl@informatik.uni-tuebingen.de>
>>>>> writes:
Martin> Section 3.3 "File system" defines a whole bunch of procedures which
Martin> take a fname/fd/port as argument and call to file-info and inspect the
Martin> file-info record (file-type...file-last-status-change,
Martin> file-directory?...file-symlink?,
Martin> file-not-readable?...file-executable). The file-info record however is
Martin> not passed to the outside so that subsequent operations on the same
Martin> file will call file-info again and again which means a huge performance
Martin> overhead.
Martin> First calling file-info and operating on the obtained record
Martin> afterwards means far less overhead but is also less comfortable since
Martin> the procedures mentioned above do not exist.
Martin> I think we can easily remedy this by allowing the mentioned procedures
Martin> to accept a file-info record as argument.
What's the good reason for overloading? Why can't we just define
additional accessors for FILE-INFO?
--
Cheers =8-} M.
Friede, Völkerverständigung und überhaupt blabla
|