Path: utzoo!attcan!cmtl01!matrox!uvm-gen!uunet!lll-winken!ames!mailrus!cornell!uw-beaver!rice!sun-spots-request From: phil@Rice.edu (William LeFebvre) Newsgroups: comp.sys.sun Subject: Re: The Moderator Always Gets the Last Word Message-ID: <2493@kalliope.rice.edu> Date: 26 Jan 89 04:50:29 GMT Sender: usenet@rice.edu Organization: Sun-Spots Lines: 48 Approved: Sun-Spots@rice.edu Original-Date: Wed, 25 Jan 89 16:58:02 CST X-Sun-Spots-Digest: Volume 7, Issue 121, message 4 of 8 This is in response to the previous article by Matt Goheen. The comment I included in John Gilmore's article was a joke. Get it? Why are so many people on the net so serious? Lighten up a little. Enjoy life. Relax! As for your nine "errors", I think you are really stretching some of them to try and say they are incorrect. > 1) "Integer overflow is (I believe) not detectable in C" The fact that a simple arithmetic operation caused the overflow bit to be set is not detectable without standard compiler or run-time library support. Yes, you can check the values before or after and determine algorithmically if the computation would cause an overflow, but you cannot actually detect the overflow event. What I'm saying is that there is nothing like a PL/I "on error goto foo;" capability. And there is no way to convert an overflow into a Unix signal: the 680x0 doesn't support that sort of behavior. > 2) The various problems with a suid uudecode program. wnl doesn't mention > the possible security hole opened when making uudecode non-suid (the > "decode" alias). There are many things that I don't mention, especially in a comment that is intended to be brief. First you tell me I put in too many comments, then you fault me for not making the comments thorough enough? You can't have it both ways, fellah. > 7) The gaffe about 8-bit characters having always been supported by the > Unix kernel, which wasn't quite true (from what I can gleam from the > responses, it wasn't properly supported until 4.3? I'm still a little > foggy on this.) Yes, this was a genuine (and rather serious) mistake. I know for an absolute and definite fact that 8-bit names were legal under 4.1. The only byte values that were not allowed were 0, '/', and '/'+0200. I cna't speak authoritatively about previous versions of Bell and BSD Unix. The change happened with 4.2 BSD and was carried over into SunOS. > 8) Not really an actual error, but mentioning the 's' vs. 'S' permissions > on group access for directories and having no idea why they were there As you said, "not really an actual error". Am I also to be faulted for admitting my ignorance? William LeFebvre