1 |
Omer Cohen posted <30e61698050422051322736ee3@××××××××××.com>, excerpted |
2 |
below, on Fri, 22 Apr 2005 14:13:51 +0200: |
3 |
|
4 |
> PHP 5.0 was released a long time ago, and alot of fixes and patches were |
5 |
> released after it to make sure it's sable. |
6 |
> |
7 |
> According to PHP.net <http://php.net/> the stable versions are PHP 5.0.4 && |
8 |
> 4.3.11 |
9 |
> 4.3.11 is marked stable, but 5.0.4 dosn't even exist on the tree. |
10 |
> |
11 |
> 5.0.0 isn't marked at all, and everything till 5.0.3-r2 is marked are hard |
12 |
> masked and still being tested. |
13 |
> |
14 |
> It's been like this for a long time now. |
15 |
> |
16 |
> As a PHP developer I believe that 5.0.4 is more then stable, and should be |
17 |
> added and marked stable. |
18 |
> |
19 |
> I don't wanna override the system and install it manualy. |
20 |
|
21 |
OK, I see a big discussion, but nobody has yet made this point, directly |
22 |
at least, so here it is... |
23 |
|
24 |
I agree that 5.0.4 should at least be in the tree, if upstream is calling |
25 |
it stable. |
26 |
|
27 |
The point that should be emphasized, however, is that there's a /big/ |
28 |
difference between the upstream application being "stable", and Gentoo's |
29 |
particular instance, that is, the ebuild script that merges it onto a |
30 |
Gentoo system, being stable. Gentoo's keywording, while somewhat |
31 |
correlating with upstream in that what upstream has declared a beta or RC |
32 |
is often never arch-stable keyworded on Gentoo, generally serves to |
33 |
indicate the Gentoo ebuild maintainer's evaluation of the stability of the |
34 |
EBUILD, *NOT* the stability (or lack thereof) of the upstream source. |
35 |
|
36 |
Thus, as I said above, yes, the version that upstream calls "stable" |
37 |
should reasonably be expected to be in the portage tree in some form |
38 |
within a reasonable (few week, often less) time, however, one can't always |
39 |
expect that said portage tree version will be marked stable just because |
40 |
upstream defines that particular version of their product as stable, |
41 |
because the status of the Gentoo instance of it, the ebuild, may itself |
42 |
not be stable, on one or more archs, possibly on all of them. |
43 |
|
44 |
In this instance, >=php-5.0 on Gentoo is hard masked, not because of what |
45 |
upsteam says, but because (presumably) there have been and remain |
46 |
unresolved issues with the Gentoo deployment. Something in Gentoo's |
47 |
previous deployments conflicts with the current 5.0 layout, and a smooth |
48 |
transition hasn't yet been worked out and fully tested, so the 5.x series |
49 |
remains hard masked. |
50 |
|
51 |
Ignoring for the moment the issue of the 5.0.4 upstream-stable version |
52 |
itself not being in the tree at all, if a sysadmin is suitably comfortable |
53 |
with php-5.x, and either understands the issues keeping it masked on |
54 |
Gentoo and knows they don't apply in his case or at least is willing to |
55 |
extend the effort to work around any issues that may appear, said sysadmin |
56 |
is entirely free to package.unmask, or add keywords in an overlay, as |
57 |
appropriate. That's why the portage system has been designed with that |
58 |
flexibility in place, after all -- so it can be used at the decision of |
59 |
the individual Gentoo user -- aka the local Gentoo system sysadmin. |
60 |
|
61 |
-- |
62 |
Duncan - List replies preferred. No HTML msgs. |
63 |
"Every nonfree program has a lord, a master -- |
64 |
and if you use the program, he is your master." Richard Stallman in |
65 |
http://www.linuxdevcenter.com/pub/a/linux/2004/12/22/rms_interview.html |
66 |
|
67 |
|
68 |
-- |
69 |
gentoo-dev@g.o mailing list |