1 |
On Saturday 29 August 2009 05:42:45 Duncan wrote: |
2 |
> Mike Frysinger posted on Sat, 29 Aug 2009 02:56:33 -0400 as excerpted: |
3 |
> > On Friday 28 August 2009 20:05:12 Alex Alexander wrote: |
4 |
> >> On Sat, Aug 29, 2009 at 00:23, Mike Frysinger wrote: |
5 |
> >> > On Friday 28 August 2009 16:27:18 Sebastian Pipping wrote: |
6 |
> >> >> Mike Frysinger wrote: |
7 |
> >> >> > 10.0 is retarded |
8 |
> >> >> |
9 |
> >> >> How would you like the problem to be addressed? |
10 |
> >> > |
11 |
> >> > we already have a simple logical version system. 2009.0 is the next |
12 |
> >> > step. |
13 |
> >> |
14 |
> >> Years do not make a good versioning scheme, if one release gets out |
15 |
> >> late you're automatically considered outdated by users. |
16 |
> > |
17 |
> > then help the release team to get more tested releases, otherwise |
18 |
> > reality is we are releasing out of date install media |
19 |
> |
20 |
> But as we all know, releases != profiles. If there's no reason to update |
21 |
> the profiles besides the fact that the name incorporates a year, and they |
22 |
> look out of date, why do so? |
23 |
> |
24 |
> For that reason, getting away from year for the profiles is a reasonable |
25 |
> idea, now that Gentoo seems to be mature enough that we don't need a new |
26 |
> profile multiple times a year. |
27 |
> |
28 |
> OTOH, having the year in there, as long as people don't get fixated on |
29 |
> it, can be useful as an indication of when the profile was born, just not |
30 |
> necessarily that it's outdated. If it weren't for the outdated |
31 |
> appearance, therefore, year would be fine. |
32 |
|
33 |
except that profiles and releases have always been tied (for good reason). |
34 |
profile default changes are made as part of the release process. if we want |
35 |
to change a USE flag default, we dont (shouldnt) be doing it to live profiles. |
36 |
it is part of the natural version bumping. releng has always been managing |
37 |
new profiles since we started the process years ago and there's no reason to |
38 |
change now. |
39 |
|
40 |
> Whatever, bikeshedding from my perspective, and this one I don't /care/ |
41 |
> what the color/name is. But since we already have 10.0 profiles in-tree, |
42 |
> just run with them, as it's more work to worry about changing them now, |
43 |
> than it's worth. (And, I might add, I'm glad they're in, as the /last/ |
44 |
> thing we need is to be stalemated debating it for a year or two, as it |
45 |
> /is/ bikeshedding.) |
46 |
|
47 |
date based profiles isnt bikeshedding, it's logical. and if your only |
48 |
complaint is that it doesnt matter, then there is absolutely no reason to go |
49 |
changing from what we've been doing for years with no complaints. picking |
50 |
random numbers out of your ass (like 10.0) is confusing. |
51 |
-mike |