Linux

Anyone familiar with it? I'm aware it's a new CVS. I've been over to the tigris website, and am considering getting a copy to play with on my Sparc Ultra II Ent. Box. Thanks in advance.


Eric

I've used it once, but decided to stick with CVS. There were a few reasons why I am sticking with CVS over subversion, but I am forgetting what those reasons were.
Subversion does have some good things going on for it, thats why I initially took a look at it, but after using it, I found it did not suit my needs.
If your starting a New project, and your in charge of the project, then I think subversion most likely will be a good choice.

Thanks, Locust. When you get a chance, I'd like to what you specifically didn't like.

In my situation, I was using CVS for a friends manuscript. Seeing that she is new to GNU/Linux in general, and I had already convinced her to use Vim and LaTeX for writing out her story, I felt that CVS would have been much more easier for her to learn.
For her, she could examine her changes easily within KDE's built-in Cervisia option.
The commands are simpler using CVS. For example under CVS, to checkout a copy of your work, you would issue;

cvs -d ~/cvsroot checkout project-files

Under subversion it is not difficult, but just longer.

svn checkout file:///home/locustfurnace/subversion/trunk project-files

Another reason was that there are more available front-ends (including Win32) for accessing CVS than Subversion. Plus, there are some filters for LaTeX for handling CVS comments.

If I recall correctly, I was not particularly a fan of the revision numbering scheme. I think subversion incremented revisions on a whole numbering system, such as 1, 2, 3, 4 while I had the ability in CVS to use 1.1.1.2, 1.1.1.3, 1.1.1.4. If necessary.

It's usually the little things which become a deciding factor for myself in what way I plan to go when picking software. The same reasons why I decided to chose LaTeX over DOCBOOK or XML when writing out a novel.
As stated earlier, if your controlling your own project, and starting from scratch. I think Subversion may have more options for you. But if your collaborating on a project, then you should factor in how those others users will be accessing the repository, and what tools they are familiar with.

Subversion is my version control system of choice for managing the LaTeX sources of a book I am working on personally, though I store most of my other scripts or Web site content in CVS server instead.

Working with CVS command line options can be a horrendous experience for new users, as option values sometimes need to be given in certain order, or that the same switch may appears twice in a command but serving different purposes. e.g.:

cvs -d :pserver:anonymous@host:/home/cvsroot co -d tutorial/ tutorial

I am impressed by Subversion for a number of reasons:
- It has less tricky dark corners compared with CVS. CVS has certain behaviour that are not intuitive unless you have read the relevant sections of the manual.
- It supports file copy/rename without losing revision history. A feature I find most useful as I don't make good filenames usually at the very start.
- It does not distinguish binary/ASCII files. Some CVS front ends are prone to always add files in binary mode, making it difficult for some CVS clients to get context diffs even though it simply involves text data.
- Repository-wide revision numbers allow state at any instant to be consistently reproduced. With CVS, you can only be reliably achieving this by tagging often. Checkout by date/time with CVS does not necessarily produce a local copy at any consistent state, depending on the frequency of commits. Per-file revision is more intuitive IMO, though.

Personally I am more used to CVS, but I appreciate SVN's new features.