Michael Jennings wrote:
Then how on Earth do you put everything right, if for some reason someone has chosen the wrong versioning scheme for earlier releases?On Tuesday, 22 February 2005, at 16:25:07 (+0100), Toralf Lund wrote:
I mean, I guess it depends on when you decide to "Say No!", but the
way I've always looked at it, epoch was pretty much designed
precisely for the event when you do say no (to weird versioning
logic.) For instance, what happens after you decide that although
the last version your little software package is 1337a.42c7, you
won't use the number normally following that (whatever that might
be) for the one you're releasing just now, but rather call it 2.0
like normal people would? Obviously, rpm won't understand that 2.0
is actually newer than 1337a.42c7, just like that. I would suggest
that what you do in this case, is to "Say No!" *and* use epoch...
Epoch is the wrong solution to a very real problem. Regardless of the
situation, just saying NO! to epoch is the right move.
And for the record, even a human couldn't tell that 2.0 is greaterI think perhaps you missed my point. Should you really do that to someone who used a ridiculous version number, but then saw the error of their ways?
than 1337a.42c7 (partly because it's *not*), so anyone who actually
used such a ridiculous version number deserves to be drug out into the
street and trampled by mousketeers.
The point was precisely that nobody could tell what is greater than 1337a.42c7. That's why switching to 2.0 (or 1.0 - but it is after all not the first release) would be a good idea. You would then go on like everyone else and call your next release 2.1, or possibly 3.0, obviously. The question is, what do you do to tell the system that you did do just that - I mean switch to a sane versioning scheme? I think the obvious answer is: You use epoch! - But you don't keep updating it, of course, as there is no need for it after introducing understandable version numbers - you just increment or add epoch at the time you release version 2.0, and keep the same value in the future.
- T
_______________________________________________ Rpm-list mailing list Rpm-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/rpm-list
