On Fri, Oct 06, 2006 at 09:28:21AM -0700, Gary Thornock wrote: > Why do I prefer CVS, then?
You have brought up some interesting points, most of which come down to a matter of personal preference. Here are some of my thoughts on these issues. > - Independent revision numbers per file, rather than a single > number for the entire project I just prefer the Subversion way. I usually look at the whole system rather than a single file. For example, when I am using a development version of a project that is kept in Subversion, I can refer to the revision number as if it were a tag. With CVS, you can't really do that. I still haven't figured out how to create a unified diff of all changes between the repository and my working tree in CVS. This is really easy with Subversion, and it works between any two revisions (or moments in time). > - Better branching and tagging For me, the way Subversion does it (making a copy-on-write copy) is really intuitive. I can see why you wouldn't like it, though. > - I've seen too many cases where Subversion's Berkeley DB > repositories have been corrupted or left in a locked state, > preventing everyone from updating or committing until an > administrator can fix things. In partial response to this > problem, Subversion has added the ability to make a > repository using the native filesystem instead of BDB, but > BDB is still the default. FSFS is the default now (and has been for at least a year). And yes, BDB is terrible. > - Subversion keeps a second copy of every file in the project > in hidden subdirectories within the working copy. Besides > being a ridiculous waste of disk space (keep in mind, I use > a laptop, so disk is at something of a premium), this also > causes problems when doing multi-file operations like > "find . -type f ..." or "foreach i (`grep -rl foo .`) ..." . > The rationale for this design was that disk is cheap and you're > not always connected to the network. I disagree. On a laptop, > disk is expensive, and between WiFi and EVDO, you pretty much > *are* always connected to the network if you want to be. Yeah. Having to add a -prune to find is a little annoying, but it's nice to be able to do some basic operations without being connected to the network. I don't have EVDO (and frankly don't know exactly what it is), and WiFi isn't widespread enough for me. Places where I have generally found terrible wireless availability include airports, airplanes, buses, parks, classrooms at BYU, classrooms at UVSC, Church buildings, etc. Basically I only get wireless at home and work and in some classrooms (although it's usually flakey). These are all relatively little points that we disagree on, but for me, CVS's complete lack of ability to rename, move, or delete files and directories is a fatal weakness. Anyway, if you are just starting with revision control and aren't already used to the way CVS does tagging and branching, I think Subversion is definitely a better way to go. -- Andrew McNabb http://www.mcnabbs.org/andrew/ PGP Fingerprint: 8A17 B57C 6879 1863 DE55 8012 AB4D 6098 8826 6868
pgpboYjjFs9Z7.pgp
Description: PGP signature
-------------------- BYU Unix Users Group http://uug.byu.edu/ The opinions expressed in this message are the responsibility of their author. They are not endorsed by BYU, the BYU CS Department or BYU-UUG. ___________________________________________________________________ List Info: http://uug.byu.edu/cgi-bin/mailman/listinfo/uug-list
