[email protected] wrote: > >You don't really describe in the bug how PUBLISHALL is broken and > >ALL_MANIFESTS is incomplete, so I don't know how to evaluate your fix. As > >far as I can tell, the only substantive change is the exit 0, but that's > >got to be a teeny corner case. > > Sorry, I realize it was a bit terse. The issue was that the shell > construct was called ALL_MANIFESTS:sh rather than the expected > ALL_MANIFESTS.<something> and then on the next line, we don't actually > end up evaluating the shell fragment at all.
I see. Does it work if you just remove the :sh from the definition of MANIFESTS? The definition of ALL_MANIFESTS is supposed to work as it is, but I can see how trying to execute its value on the next line would give you crud. > So the fix is to rename > the construct along the lines of the other cases in the Makefile and > then properly evaluate it using $(ALL_MANIFESTS.cmd:sh). > > ALL_MANIFESTS.cmd = \ > . > . > . > ALL_MANIFESTS = $(ALL_MANIFESTS.cmd:sh) > > To be consistent with the other examples, I also broke this up into > separate lines. Okay. And while you're changing i386_DEFINES and sparc_DEFINES, you might as well update line 92 (ARCH_DEFINES) to use the same whitespace pattern as everything else. > >Also note that there's no need to put diffs into the suggested fix. That's > >why we put changeset ids into the bug. > > Aye, just habit out of using Bugster. Understood. We should stop doing this in ON, too. Anyway, +1 on what you've got. Thanks, Danek _______________________________________________ pkg-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
