| 1 |
Jurek Bartuszek <jurek@g.o> posted 462E71E1.7000704@g.o, |
| 2 |
excerpted below, on Tue, 24 Apr 2007 23:08:49 +0200: |
| 3 |
|
| 4 |
>> Existing _rcX cases can be handled like this: |
| 5 |
>> |
| 6 |
>> _rc2-rYYYYMMDD |
| 7 |
>> |
| 8 |
>> Portage will update from _rc2 to a version with revision part > 0. |
| 9 |
> |
| 10 |
> However, _rc2-rYYYYMMDD-r1 would *not* be valid anymore, and I think |
| 11 |
> it's quite easy to imagine when this additional -r1 would be neccessary. |
| 12 |
|
| 13 |
Well, since -rX is revision number, what was proposed is in effect using |
| 14 |
a dated revision number, so instead of _rc2-rYYYYMMDD-r1, it would be a |
| 15 |
new -rYYYYMMDD. The additional data that the second one is simply a |
| 16 |
revision of the first would of course be lost. |
| 17 |
|
| 18 |
OTOH, what happens then when another snapshot is taken but not yet |
| 19 |
stabilized, and say a security revision of the first snapshot is |
| 20 |
required. Then we have two different -r<date> snapshot sequences |
| 21 |
interleaving. |
| 22 |
|
| 23 |
So then to cure that we end up with this: |
| 24 |
|
| 25 |
_rc2-rYYYYMMDDrr, where rr being two digits taking the place of the |
| 26 |
second revision sequence ( the -rX in _rc2-rYYYYMMDD-rX) in the example |
| 27 |
above. Human parsing of that long a string of digits becomes |
| 28 |
increasingly difficult, unfortunately, but it should work. |
| 29 |
|
| 30 |
-- |
| 31 |
Duncan - List replies preferred. No HTML msgs. |
| 32 |
"Every nonfree program has a lord, a master -- |
| 33 |
and if you use the program, he is your master." Richard Stallman |
| 34 |
|
| 35 |
-- |
| 36 |
gentoo-dev@g.o mailing list |
Archives