From owner-freebsd-hackers Mon May 25 03:17:22 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id DAA16128 for freebsd-hackers-outgoing; Mon, 25 May 1998 03:17:22 -0700 (PDT) (envelope-from owner-freebsd-hackers@FreeBSD.ORG) Received: from ns1.yes.no (ns1.yes.no [195.119.24.10]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id DAA16120 for ; Mon, 25 May 1998 03:17:17 -0700 (PDT) (envelope-from eivind@bitbox.follo.net) Received: from bitbox.follo.net (bitbox.follo.net [195.204.143.218]) by ns1.yes.no (8.8.7/8.8.7) with ESMTP id KAA06497; Mon, 25 May 1998 10:17:12 GMT Received: (from eivind@localhost) by bitbox.follo.net (8.8.8/8.8.6) id MAA02033; Mon, 25 May 1998 12:17:03 +0200 (MET DST) Message-ID: <19980525121702.27321@follo.net> Date: Mon, 25 May 1998 12:17:02 +0200 From: Eivind Eklund To: Studded , freebsd-hackers@FreeBSD.ORG Subject: Re: [Fwd: Intermediate (?) level CVS questions] References: <356485A5.8022EEE2@san.rr.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.89.1i In-Reply-To: <356485A5.8022EEE2@san.rr.com>; from Studded on Thu, May 21, 1998 at 12:51:01PM -0700 Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG On Thu, May 21, 1998 at 12:51:01PM -0700, Studded wrote: > First question. I have successfully set up the CVS repository and used > it to check out a full set of -Stable sources, however there are quite a > few items that I don't build even during a make world so I would like to > avoid checking them out to conserve disk space. I looked high and low > for a setting that says "check out all of 'src' except the following" > similar to the refuse file I use for cvsup but couldn't find it. I've > basically come to the conclusion that such an option doesn't exist, > which isn't the end of the world I suppose. I know I could make a script > that checks out the individual elements that I want (in fact I did that > for ports) but it would be more work than I want to keep that up to date > for the src tree, and work is what I'm trying to avoid with CVS. :) You might be able to use .cvsignore file to avoid having this checked out, I think. (I think the behaviour here might differ from release to release of CVS, but I don't really remember). If it presently doesn't work, it would probably be an easy change to add. > Related to the first question, I have customized a lot of stuff in my > make world setup. Almost all of these changes are related to parts of > the system that I don't build, so we're essentially talking about a lot > of modified makefiles, and a few things related to long usernames. What > I'd like to do is set up a branch that has all of my stuff on it so when > I check out the src tree I can merge my branch into the main branch. I > am familiar with creating branches, however I ran into two problems. > First, I edited one of the makefiles, tagged it, then tried 'cvs commit > -r MYBRANCH Makefile'. I got an error message asking me to commit on > freefall instead which I'm sure no one would want me to do even if I > could. :) However I think this point is moot because I figured that > every time I cvsup'ed the cvs tree it would delete my changes, and I was > correct in that when I tested. You have to use 'magic branches' for this. Basically, you will want to have your branch start at a sub-revision of greater than 100 - ie, for a branch 1.X.Y you want Y to be 100 or more. However, I'm not certain how useful this will be. > What I suspect is that I need to create a whole new directory under > /home/ncvs for things like this, but I'm befuddled on exactly how to do > that from things checked out to /usr/src. I'd also like to do something > similar here with things like CAM. This is one area where I could use > some specific, detailed help or pointers to the same. :-/ You can't do things like this easily. It is beyond the capabilities of CVS. > The next problem I'm having is using the rdiff command. One of the main > reasons I want the full CVS repository on hand is to take a look at the > changes between 2.2.5 and 2.2.6 and see if I can find the cause of some > performance problems one of my customers is having since the upgrade. > When I do: > > cvs rdiff -kk -r RELENG_2_2 -r HEAD src > > I get a nice fat diff file and all kinds of little updates on the screen > telling me where cvs is diff'ing. However when I do: > > cvs rdiff -kk -r RELENG_2_2_5_RELEASE -r RELENG_2_2_6_RELEASE src > > I get bupkis. Nothing but a grinding disk and no output at all. I've > double checked the tag names but still can't figure out where I'm going > wrong. I can't tell why this would be a problem; it worked for me. (Breaking it exposed recursive use of malloc() in a signal handler in cvs, but this absolutely doesn't produce "No output at all" - I've not seen that much output for quite a while ;-) Eivind. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message