On Sun, Jun 12, 2005 at 10:26:19PM -0400, Roberto C. Sanchez wrote: > On Sun, Jun 12, 2005 at 08:12:46PM -0600, Paul E Condon wrote: > > > > It purports to be information that is available to script writers > > when, if fact, it is not information. It does not, in and of itself, > > cause a problem. It just misleads script writers into believing that > > there is a really simple way to determine the Debian version, when in > > fact there is not. At least not by reading this file. It would not be > > a bug if it were totally undocumented and given another name so that > > script writers who notice it would not assume falsely that it is > > useful. > > > > I disagree. This file is *guaranteed* to have different contents for > every stable release. Using that file I instantly tell the difference > between a Potato, Woody and Sarge install. I just did a dist-upgrade of a spare box from Sarge to Etch. The contents of /etc/debian-version was 3.1 before the dist-upgrade and remain the same afterwards. Etch, of course, has no version number. So there is no way to test a script that has version dependency until it is released into stable. This is, I think, a very strange way to do business.
I conclude that all knowledgeable Debian developers do not use the contents of /etc/debian-version. I don't know how each new Debian developer first learns about this trap, but it is a trap, and should be eliminated, IMHO. -- Paul E Condon [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
