1 |
On Wed, 2006-09-13 at 07:52 -0700, Brian Harring wrote: |
2 |
> On Wed, Sep 13, 2006 at 10:34:52AM -0400, Aron Griffis wrote: |
3 |
> > From bind-9.3.2-r4.ebuild: |
4 |
> > |
5 |
> > # idea from dev-libs/cyrus-sasl |
6 |
> > if has distcc ${FEATURES}; then |
7 |
> > einfo "You have \"distcc\" enabled" |
8 |
> > einfo "build with MAKEOPTS=\"-j1\"" |
9 |
> > jobs="-j1" |
10 |
> > else |
11 |
> > einfo "build with MAKEOPTS=${MAKEOPTS}" |
12 |
> > jobs="" |
13 |
> > fi |
14 |
> > |
15 |
> > emake ${jobs} || die "failed to compile bind" |
16 |
> > |
17 |
> > I think this is bogus. If building with distcc reveals a parallel |
18 |
> > build issue, then the issue exists with or without distcc, it just |
19 |
> > seems to happen less often without it. We've been down this road |
20 |
> > before, maybe people have forgotten? |
21 |
> > |
22 |
> > bind-9.3.2-r4.ebuild failed to build for me on dual ia64. Building |
23 |
> > with -j1 works. |
24 |
> > |
25 |
> > Unless somebody can explain how this is valid, I'll go ahead and fix |
26 |
> > the bind ebuilds (where "fix" means "use -j1 unconditionally since the |
27 |
> > Makefiles aren't parallel safe"). |
28 |
> |
29 |
> Similar trickery in app-office/openoffice, although they enable -jN if |
30 |
> distcc is enabled, else -j1 ... |
31 |
> |
32 |
> Always wondered how that was valid, just avoid OO compiles enough it |
33 |
> wasn't something I ever got around to looking into :) |
34 |
> ~harring |
35 |
|
36 |
I don't see how it can be valid, especially ferringb's example. |
37 |
Enabling distcc doesn't mean the build will distribute; only that the |
38 |
possibility is there. If you happen to build when the other system(s) |
39 |
is(are) unavailable for some reason, you get everything on the host, so |
40 |
for ferringb, e.g., this would mean -jN on one system. |
41 |
|
42 |
As for Aron's case, I agree with him. |
43 |
|
44 |
Regards, |
45 |
Ferris |
46 |
-- |
47 |
Ferris McCormick (P44646, MI) <fmccor@g.o> |
48 |
Developer, Gentoo Linux (Devrel, Sparc) |