| 1 |
On Sun, 17 Jul 2011 11:19:31 +0200 |
| 2 |
justin <jlec@g.o> wrote: |
| 3 |
|
| 4 |
> On 17/07/11 11:16, Michał Górny wrote: |
| 5 |
> > On Sun, 17 Jul 2011 11:05:39 +0200 |
| 6 |
> > justin <jlec@g.o> wrote: |
| 7 |
> > |
| 8 |
> >> Hi all, |
| 9 |
> >> |
| 10 |
> >> the following change has some bad effects on packages in the tree. I |
| 11 |
> >> found it with a package having this inherit line |
| 12 |
> >> |
| 13 |
> >> inherit cmake-utils wxwidgets |
| 14 |
> >> |
| 15 |
> >> all cmake-utils_* function which have redundant phase definitions in |
| 16 |
> >> base will not be executed. |
| 17 |
> >> |
| 18 |
> >> This change was not trivial. So either we need to revert it or really |
| 19 |
> >> check the tree that nothing is broken. |
| 20 |
> > |
| 21 |
> > Revert. The wxwidgets eclass is not really supposed to export phase |
| 22 |
> > functions, I've seen packages in Sunrise using it along with |
| 23 |
> > cmake-utils too. If someone wants user patches in their package, he/she |
| 24 |
> > should inherit base directly and get phase funcs clean. |
| 25 |
> > |
| 26 |
> |
| 27 |
> So I checked it. There are a couple of packages which use it in |
| 28 |
> combination with cmake-utils, python and autotools-utils. And as |
| 29 |
> maintainer keeps the lines sorted, all those ebuilds are broken potentially. |
| 30 |
|
| 31 |
And people wonder why I hate eclasses... |
| 32 |
|
| 33 |
|
| 34 |
-- |
| 35 |
fonts, gcc-porting, it makes no sense how it makes no sense |
| 36 |
toolchain, wxwidgets but i'll take it free anytime |
| 37 |
@ gentoo.org EFFD 380E 047A 4B51 D2BD C64F 8AA8 8346 F9A4 0662 |
Archives