[PR40159,trunk,vta,vta4.4] don't succeed on failures in recursive make

[Dave Korn]

Alexandre Oliva wrote:
> On May 18, 2009, Alexandre Oliva <aoliva@redhat.com> wrote:
> 
>> Andreas Schwab reminds me I'm no longer suited to be a configury
>> maintainer :-)
> 
> And H-P reinforces the point ;-) reminding me gcc-bootstrap and
> gcc-no-bootstrap aren't mutually exclusive.  :-/

  I hate to make you feel like it's beat-up-on-Alex day :-) but I wanted to
remind you:

 @if gcc-bootstrap
-	if [ -f stage_last ]; then : ; \
+	if [ -f stage_last ]; then \
 	  TFLAGS="$(STAGE$(shell sed s,^stage,, stage_last)_TFLAGS)"; \
 	  $(MAKE) $(TARGET_FLAGS_TO_PASS) all-host all-target; \
 	else \
 @endif gcc-bootstrap


... $(shell) is evaluated by make, regardless of the [ -f ] test.  This leads
to a harmless (I think?) warning at the start of a clean build:

$ make -j4 2>&1 | tee build-parallel.log
sed: can't read stage_last: No such file or directory
[ -f stage_final ] || echo stage3 > stage_final

    cheers,
      DaveK