Hello,
On 30/09/2011 08:51, Ian Campbell wrote:
What does the cleanup actually do?
IIRC, the patch meant to do 2 things:
* A small improvement (in two occasions)
E.g., for make rule
lib$(1)_stubs.a: $(foreach obj,$($(1)_C_OBJS),$(obj).o)
substitute the action part from
$(call mk-caml-lib-stubs,$$@, $$+)
to
$(call mk-caml-lib-stubs,$(1), $$+)
i.e. pass part of the target's name as the argument to mk-caml-lib-stubs
instead of the whole name, so that we can avoid calling basename and sed later
on to recover the sub string. I.e.,
from
$(call quiet-command, $(AR) rcs $1 $2 && $(OCAMLMKLIB) -o `basename $1
.a | sed -e 's/^lib//'` $2,MKLIB,$1)
to
$(call quiet-command, $(AR) rcs lib$(1)_stubs.a $2 && $(OCAMLMKLIB)
$(LIBS_$(1)) -o $(1)_stubs $2,MKLIB,$1)
* Some correction on linking parameters for bytecode mode compilation
IIRC, it's this line that changed from
$(call mk-caml-lib-bytecode,$$@, -dllib dll$(1)_stubs.so -cclib
-l$(1)_stubs, $$+)
to
$(call mk-caml-lib-bytecode,$$@, -dllib dll$(1)_stubs.so -cclib
-l$(1)_stubs $(foreach lib,$(LIBS_$(1)),-cclib $(lib)), $$+)
The former was able to produce bytecode as well, but there would be problem in
dynamic loading the generated bytecode (e.g. in OCaml toplevel). The later
version should work fine with the toplevel.
-LIBS_evtchn = $(LDLIBS_libxenctrl)
+LIBS_eventchn = -L$(XEN_ROOT)/tools/libxc -lxenctrl
This should continue using LDLIBS_libxenctrl unless there is a good
reason not too.
I think you are right. They are identical, only the name in the first line
needs fixing, and using predefined $(LDLIBS_libxenctrl) is certainly better.
The change dated back before your Makefile refactoring after Xen-4.1 and was
probably a result of mix up during rebase.
Cheers
--
Zheng
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel
|
Home