* John Hasler (john@dhh.gt.org) [010810 08:00]: > David Roundy writes: > > I just checked, and ls reports that the file is empty (no great > > surprise). So presumably the latter editors first check the size before > > reading the file, while the former ones simply read from it until they > > hit EOF. > > Yes, that makes sense. A very minor kernel bug. I don't know that it's fair to call this a bug. Consider the alternative: each time a directory in /proc is accessed the kernel creates all the output necessary for each "file" therein to determine the length, just to be able to fill in correct values for ls? Even still, it would be broken, as stuff in /proc can change instantly, so the lengths returned would only be ephemerally correct anyway. It's better this way. It hardly makes sense to try to $EDITOR these "files" anyway; just use echo and cat. (You're usually just reading or writing one byte anyway). -- Vineet http://www.anti-dmca.org Unauthorized use of this .sig may constitute violation of US law. Qba\'g gernq ba zr\! |tr 'a-zA-Z' 'n-za-mN-ZA-M'
Attachment:
pgp6cKz4y1JMM.pgp
Description: PGP signature