[cvsnt] Re: "cvs commit -r " problem

[Julian Opificius]

Rick Genter wrote:
>>
>>Tags? You're joking, right?
> 
> 
> Yes and no. For the scenario cited (code review), I've found that
> usually a code review is a planned event and not intended to be
> performed on an arbitrary revision of code. In fact, in code review
> scenarios I've been in, we've usually wanted to review the code that
> went into a particular QA build, for example, for which a tag does
> exist.
> 
Good point. I do the same in that scenario.
> 
>>Why in the heck would I want to tag every single incremental revision
> 
> of 
> 
>>a file when the automatically assigned revision number is a perfectly 
>>valid way of doing so?
> 
> 
> I have no problem with using an automatically assigned revision number.
> What started this thread was someone wanting to override the
> automatically assigned revision number. That's something that I (and the
> CVSNT developers) have a problem with. If you want something other than
> the automatically assigned revision number, use a tag.
>

I agree with you - strongly, in fact: the ability to force revision 
numbers would make it very difficult for me to justify the use of CVS or 
CVSNT for FAA work.

I just thought for a moment you were proposing that any access to a 
non-head revision would need a tag, and that didn't seem sensible.

Again, though, I'm worried that any new internal revision identification 
mechanism might lose the inherent chronological implications of the 
current numerical system.

j.