On Mon, May 20, 2002 at 11:34:01PM +0100, Colin Watson wrote: > On Mon, May 20, 2002 at 10:30:28PM +0200, Jeroen Dekkers wrote: > > On Sun, May 19, 2002 at 07:11:57PM -0700, Thomas Bushnell, BSG wrote: > > > Now if you want to point out that FHS doesn't mention /hurd and so > > > using it is in violation of FHS, then you are quite right. > > > > No, then he's wrong. I don't know about your copy of the FHS, but mine > > (version 2.2 dated May 23, 2001) doesn't forbid a distribution to add > > directory. > > Looking at version 2.2 at http://www.pathname.com/fhs/, section 3.1, it > explicitly says that "software must never create or require special > files or subdirectories in the root directory", over and above those > specified in the standard. As a result, my reading suggests that > distributions including such software aren't compliant with the current > version of the FHS. Which part are you looking at? "Software must never create or require special files or subdirectories in the root directory. Other locations in the FHS hierarchy provide more than enough flexibility for any package." 1) You have to quote the whole text. This is clearly about third-party packages and doesn't say anything about distributions. 2) This is part of the rationale. A rationale only clarifies things and is not really part of the standard. But the FHS people are defining things in the rationale and clarifying things in non-rationale sections. It's a hell if you want to check whether a given system is conforming or not. > > It's just interesting to see that nobody has even noticed this, but of > > course I'm the clueless flamebait here, it's just that I have read the > > FHS carefully more than once while writing that GNU/Hurd specific > > annex and checking where Debian GNU/Hurd differs from the FHS at the > > moment. It's not that I don't know where I am talking about. > > > > I suggest people should read it carefully before trying to discuss > > these things and then they will even find out that it's amateurish and > > confusing like hell. Don't try to even try to reply without reading it > > carefully as I'm not going to waste time anymore teaching somebody to > > read. > > Notwithstanding this, and I have read it somewhat carefully, I'd be > interested to hear chapter and verse. It's clear that you have a > particular piece of text in mind. (I have no interest in bashing the > Hurd for not conforming to a document that clearly hasn't taken account > of it yet, by the way - that would be pointless and, as tb has pointed > out, off-topic.) I said *carefully*. But I agree this standard is confusing like hell. > > Keeping in mind that the FHS is a document written for proprietary > > third-party LSB packages and it specifies what those packages can find > > where and where they should install things helped me a lot with > > understanding it. > > The FHS isn't the LSB, and it certainly claims (1.1) to be intended for > more than just creators of third-party packages. Debian doesn't use it > in the sense you say, as far as I know. The authors can say anything, that doesn't matter. Linus also said he was going to write a whole operating system, I've never seen him doing that. > > Anyway, it's just useless trying to convince people on debian-devel > > that they are wrong because they don't want to listen anyhow. > > I'm quite prepared to admit I'm wrong, but I'd like to hear references. > Are you? I'm prepared to admit that I'm wrong when I'm wrong, not when 20 people say I'm wrong. But do you think the FHS is a standard that can be easily understand? I don't think so and I blame the authors for that. Jeroen Dekkers -- Jabber ID: jdekkers@jabber.org IRC ID: jeroen@openprojects GNU supporter - http://www.gnu.org
Attachment:
pgpvvLnQ_Bpay.pgp
Description: PGP signature