| 1 |
"Mark Knecht" <markknecht@×××××.com> posted |
| 2 |
5bdc1c8b0811231435y2eea1b1by366c787c983c1089@××××××××××.com, excerpted |
| 3 |
below, on Sun, 23 Nov 2008 14:35:42 -0800: |
| 4 |
|
| 5 |
> The root cause of this is I wanted to emerge the newest version of |
| 6 |
> Ardour from the pro-audio overlay and ran into a new message about it |
| 7 |
> being masked by something called EAPI 2 which according to the message |
| 8 |
> requires a 'newer' version of portage. (No revision given.) That's all |
| 9 |
> this was about, and there's absoutely no rush to fix it as I'm not |
| 10 |
> likely to really use Ardour. Just wanted to take a look at what sort of |
| 11 |
> headway they are making with their feature set so it's really nothing |
| 12 |
> but pure curiosity. |
| 13 |
|
| 14 |
OK. FWIW, for EAPI-2, you need the new ~arch portage-2.2-rcX series. |
| 15 |
EAPI-2 is allowed in various overlays and now in ~arch, but not in stable |
| 16 |
until a stable portage supports it. It'll bring a number of new features |
| 17 |
including full set support, per-package use-defaults (previously a USE |
| 18 |
flag could be defaulted to on per profile, but not per package, off being |
| 19 |
the unset default, of course), and IIRC use dependencies (if a package |
| 20 |
requires say C++ support and gcc has been built without it, it must now |
| 21 |
die with an error message telling the user to make the change, with use- |
| 22 |
deps, it could force gcc to be recompiled with C++ instead of dying, thus |
| 23 |
avoiding somebody leaving a 200-package emerge going overnight, only to |
| 24 |
come back the next day to find out it stopped with package #2 due to a |
| 25 |
USE dependency death). |
| 26 |
|
| 27 |
So there are some nice things coming in EAPI-2 and a number of packages |
| 28 |
can really use them. But an EAPI-2 supporting portage, while now in the |
| 29 |
tree, remains unstable, as there are still a few bugs to work out before |
| 30 |
it goes fully stable. So if you prefer a stable portage, you'll have to |
| 31 |
wait for EAPI-2, and any packages requiring it (which by definition can't |
| 32 |
be stabilized until an EAPI-2 portage is stable too). |
| 33 |
|
| 34 |
-- |
| 35 |
Duncan - List replies preferred. No HTML msgs. |
| 36 |
"Every nonfree program has a lord, a master -- |
| 37 |
and if you use the program, he is your master." Richard Stallman |
Archives