1 |
On Wed, 07 Nov 2012 16:05:25 Jeremy Olexa wrote: |
2 |
> On Wednesday, November 7, 2012, Francois Bissey wrote: |
3 |
> > On 07/11/12 21:08, Konstantin Tokarev wrote: |
4 |
> > > 07.11.2012, 07:03, "François Bissey" |
5 |
> > > <francois.bissey@×××××××××××××.nz<javascript:;>> > |
6 |
> > >> Hi, |
7 |
> > >> |
8 |
> > >> I have currently a prefix on OS X and I use gcc-apple. I tried a couple |
9 |
> > >> of times to use vanilla gcc which is currently masked but I had |
10 |
> > |
11 |
> > problems. |
12 |
> > |
13 |
> > > You can unmask it and emerge, than use it as your default portage |
14 |
> > |
15 |
> > compiler. |
16 |
> > |
17 |
> > |
18 |
> > Yes and it will regularly fail to compile stuff. I have tried it this |
19 |
> > afternoon and it failed to compile glib (that was gcc-4.6.3) it is |
20 |
> > masked for a reason. I am just wondering what we want to do long term |
21 |
> > because vanilla gcc seem to be unreliable. |
22 |
> |
23 |
> "We" do nothing. You can poke upstream of your favorite project to make the |
24 |
> software work with older compilers or provide patches. It is just |
25 |
> unreasonable to ask the downstream enabler to fix random projects. |
26 |
> -Jeremy |
27 |
> |
28 |
|
29 |
I am still unsure if Jeremy was trolling basically saying that software should |
30 |
support ancient compilers and basically forgo new features or may be have |
31 |
a fall back (like that will happen especially with C++). |
32 |
Anyway I synced yesterday and the last victim is eix. I remember the days |
33 |
before eix - painfully. |
34 |
CXX=clang++ emerge -uv eix worked well but using gcc-apple failed miserably. |
35 |
|
36 |
A further question: Is it, in the interim of a better solution, OK to depend |
37 |
and set the compiler to clang on macos in a ebuild or should we just put a |
38 |
warning that they will have to set clang in pkg-setup? |
39 |
|
40 |
Francois |