Re: Extending version numbering (Was: glibc_2.0.7pre1-3)

["Adam P. Harris" <apharris@onshore.com>]

[You (Rob Browning)]
>Yann Dirson <ydirson@a2points.com> writes:
>> The problem I have with epochs is that it will have to bump each time
>> a packages enters a "pre" period.  I fell that's not very nice.
>
>What difference does it make?

In truth, non.

OTOH, I too would be very hesitant to instance an epoch, given that I'm 
stuck with that epoch always and forever.

It would be theoretically possible, and I think, enormously desirable to
have some sort of "sub-epoch", say call it 'X-Y:<ver>', such that it
overrides only Y subversions into the upstream version:
    2.0.8-1 greater-than 1-3:2.0.7-1 greater-than  2.0.7pre1

Do you see what I'm getting at?  A standard epoch could in fact be 
represented as, i.e., '1-0:', that is, overriding in every case, and 
could have the short-hand notation of the epoch we currently have, i.e., 
'1'.

Problems:
  * even uglier than epoch, but who cares
  * wouldn't help us get rid of current epoch, but who cares

Anyone like this?

.....A. P. Harris...apharris@onShore.com...<URL:http://www.onShore.com/>