Numbering for cvs-cur

[Stephen McKay]

On Monday, 2nd June 2003, "Poul-Henning Kamp" wrote:

>In message <3EDBA820.8070703 at math.missouri.edu>, Stephen Montgomery-Smith writes:
>>A while back, someone (I forget who) suggested that I simply try the 
>>software out to see what would happen when the 4 digits run out for 
>>numbering the cvs-cur deltas.  Well, I am happy to report that the 
>>programs seem to seemlessly switch to 5 digits.  You should see this 
>>happen with cvs-cur sometime at the end of this year.  The change should 
>>be largely transparent.  (If a problem should later arise, I will deal 
>>with it then, probably by restarting at 100.)

Ah, that would be me.  I also read the code in case it was working
just by chance (by overflowing buffers or similar) and my expectation
is we should be able to go to about 10 digits, ie the limits of a
signed 32 bit integer.  Enough, I think. :-)

>You guys have _no_ idea how amazed I am that this would ever become
>an issue :-)

It's hard to explain but it's just so handy having the source plop into
my mailbox, even though it's been a long time since I've had to use a
dialup modem.

Comfy like an old shoe! ;-)

Stephen.