I wrote:
> Hm ... you're cd'ing into src/pl/plpython and issuing "make all"?
> That works for me.
> ... or, wait ... with -j it doesn't.  That's strange, will look.

So after a bit of digging, it seems that the locution

all: submake-generated-headers

doesn't result in ensuring that submake-generated-headers is complete
before we go to build the other targets required by "all"; it only
says that submake-generated-headers must be complete before we execute
the (empty) list of commands attached to the "all" target.

I'd tested high-j runs pretty carefully at top level, but it turns
out that that works because in both the toplevel GNUmakefile and
src/Makefile, all the interesting work happens in recursive sub-makes,
and we force the ordering of those properly with the dependencies on
the recursive make rules:

$(1)-$(2)-recurse: $(if $(filter all install, $(3)), submake-generated-headers) 
$(if $(filter check, $(3)), temp-install)
        $$(MAKE) -C $(2) $(3)

If you go to, eg, src/pl/plpython and issue "make -j", there's nothing
to prevent the builds of object files from happening before the header
build finishes.  There *was* something there before:

$(OBJS): | submake-generated-headers

but I took it out thinking it was no longer needed.

The short-term solution seems to be to put that back, but that's sort
of annoying because it means this isn't a bulletproof solution.  It
will only work for builds started in one of the directories that we
take the trouble to put this defense into, and I can't see doing that
everywhere.  Still, such things didn't work reliably before either
except in these few directories, so maybe it won't matter.

                        regards, tom lane

Reply via email to