The file format statement wasn't mine, but I would assume that the inference is that the stack metaphor is not modular enough to support a control-level current versioning system. I don't know whether or not that is true, but it certainly isn't lending itself to it from where I sit.
Basically, my point was that the binary file is not the best when used with CVS, since you cannot track changes with any specificity. CVS is uses the "diff" program, which deals with changes on a text line by text line basis, and is quite smart at spotting insertions and deletions. CVS stores the final version in the repository file, along with diffs to each previous version. For binary, you have to tell CVS it is binary, and it will just store each version verbatim. With text files, you can ask CVS to tell you which lines were changed in the last checkin from any previous checkin, which is very useful if you are trying to track down a bug. It is also useful for tracking worker productivity (although a small amount of code that works is a lot better than a lot of code that doesn't, it is unfortunately a necessary management function to track progress in some way or another.)
-Steve
_______________________________________________ use-revolution mailing list [EMAIL PROTECTED] http://lists.runrev.com/mailman/listinfo/use-revolution
