Date: Tue, 27 Feb 1996 18:33:44 -0500 (EST) From: Chuck Robey <chuckr@glue.umd.edu> To: Peter Wemm <peter@jhome.DIALix.COM> Cc: Bruce Evans <bde@zeta.org.au>, CVS-committers@freefall.freebsd.org, cvs-all@freefall.freebsd.org, cvs-lib@freefall.freebsd.org, pst@freefall.freebsd.org Subject: Re: cvs commit: src/lib/libc/db - Imported sources Message-ID: <Pine.OSF.3.91.960227182920.13888C-100000@thurston.eng.umd.edu> In-Reply-To: <199602271631.AAA20908@jhome.DIALix.COM>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, 28 Feb 1996, Peter Wemm wrote: > >> src/lib/libc/db - Imported sources > >> Update of /home/ncvs/src/lib/libc/db > >> In directory freefall.freebsd.org:/a/pst/db.1.85 > >> > >> Revision/Branch: 1.1.1 > >> > >> Log Message: > >> Import updated Berkeley DB into CSRG branch > > > >This breaks everything that hasn't moved off the CSRG branch. > >E.g., the current btree.h changed and it is now inconsistent with > >the current bt_close.c which didn't change. > > > >Bruce > > The include/db.h and include/mpool.h update was botched too. I'm a few > minutes away from finishing a 'make world' after attempting to repair it. I'll > commit it in a moment. > > Just a reminder for everybody out there.. Importing stuff has potentially > serious consequences if you are not aware of what is going to happen. Peter, I'm learning here. You are clear about vendor branch, does that mean if the old db code is in the main libc, and it wasn't changed, that changing the CSRG libc code will change the code for -current's libc? I'm concerned because I know the db is used for passwords, and that the old db code (I'm pretty sure) makes db files that aren't compatible with db.1.85's files. I don't want to do a make world, log out, and find I'm shut out of my machine. > > If a file was imported back when the 4.4Lite tree was built and NEVER touched, > it is "still on the vendor branch". Ie: it is checked out with revision > 1.1.1.1 or 1.1.1.2 etc. If you import a new version into the tree, the vendor > branch is updated, and all the files that are still connected to it > spontaniously updated. > > However, If a file has left the vendor branch and started on the mainline, > doing an import only updates the vendor branch. You have to go to each and > every changed file and manually merge the changes onto the mainline. > > What Paul has done is import new code onto the vendor branch. 90% of the db > files were still on the vendor branch, and 10% had been modified (whitespace, > fixes, and thread-safe updates). Those 10% of the files were still at the > db.1.73 level, while the unmodified 90% were now at db.1.85 level. > > The internal interfaces have changed significantly, preventing a compile. > > The entire kernel is off the vendor branch, so imports there are harmless. > However, 90% of the user-mode code is still at 4.4Lite vendor level. This is > why we cannot "trivially" import lite-2. It would take a week to get the tree > to compile again and probably longer to fix the mysterious problems. > > And what Paul did in the include directory is *definately* off limits. > Committing onto the vendor branch directly will cause me to dust off the > baseball bat that I inherited from Rod Grimes, and start aiming for people's > kneecaps. The "only" way to get stuff on the vendor branch is with 'cvs > import'. Fair warning to all... > > (Nothing personal against Paul here.. I'm merely stating the position I intend > to take) > > Cheers, > -Peter > ----------------------------+----------------------------------------------- Chuck Robey | Interests include any kind of voice or data chuckr@eng.umd.edu | communications topic, C programming, and Unix. 9120 Edmonston Ct #302 | Greenbelt, MD 20770 | I run Journey2 and n3lxx, both FreeBSD (301) 220-2114 | version 2.2 current -- and great FUN! ----------------------------+-----------------------------------------------
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.OSF.3.91.960227182920.13888C-100000>