| 1 |
Michael Palimaka posted on Fri, 26 Jul 2013 23:36:59 +1000 as excerpted: |
| 2 |
|
| 3 |
> On 26/07/2013 04:26, Duncan wrote: |
| 4 |
>> but that doesn't take care of ebuilds hard-enabling build-time deps, |
| 5 |
>> which then cause the build to fail when they're not found. Those hard |
| 6 |
>> enablings must be patched out, and if that's being done, might as well |
| 7 |
>> patch out the dependency itself at the same time. |
| 8 |
|
| 9 |
> Which ebuilds hard-enable? |
| 10 |
|
| 11 |
As I've said elsewhere just now, wrapping up threads long overtaken by |
| 12 |
events that I still had marked to reply to later... |
| 13 |
|
| 14 |
In context, I was referring to (the now dead issue of) gentoo/kde |
| 15 |
removing USE=semantic-desktop. While the USE flag was gone, ebuilds that |
| 16 |
had previously soft-depended on nepomuk, etc, subject to USE=semantic- |
| 17 |
desktop, were then hard-depending on it, because the USE flag had been |
| 18 |
removed and the dependencies hard-enabled in the ebuild. |
| 19 |
|
| 20 |
Gentoo's package.provided could be used to "fake" the package being there |
| 21 |
for portage, but that wouldn't help for ebuilds that hard-enabled |
| 22 |
configure-options that had previously been enabled only with USE=semantic- |
| 23 |
desktop, because the dependencies were now hard-dependencies coded into |
| 24 |
the ebuild. Naturally, when those ebuilds failed to find dependencies |
| 25 |
the hard-enabled options called for, they failed, and package.provided |
| 26 |
wouldn't help with that. |
| 27 |
|
| 28 |
I was simply saying that such hard-enabling had to be patched out to |
| 29 |
avoid those failures, and since we were patching it out anyway, we might |
| 30 |
as well patch out the entire dependency, thus avoiding the whole |
| 31 |
package.provided hassle as well. |
| 32 |
|
| 33 |
Happily, events overtook the thread in my absence, and gentoo/kde decided |
| 34 |
to bring back USE=semantic-desktop before 4.11 stabilization, after all. |
| 35 |
=:^) So now it doesn't matter. However, that /is/ what I was referring |
| 36 |
to, as I wrap up this subthread. |
| 37 |
|
| 38 |
-- |
| 39 |
Duncan - List replies preferred. No HTML msgs. |
| 40 |
"Every nonfree program has a lord, a master -- |
| 41 |
and if you use the program, he is your master." Richard Stallman |
Archives