This is the mail archive of the libc-alpha@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [RFC 1/1] testcase fail while using cross-test-ssh wrapper


On Mon, Jul 08, 2019 at 02:09:13PM +0200, Florian Weimer wrote:
> * Mao Han:
> >  
> >  common_objpfx=$1
> > -test_wrapper_env=$2
> > -run_program_env=$3
> > +test_wrapper_env=$3
> > +run_program_env=
> >  
> >  LIBPATH="$common_objpfx"
> 
> Sorry, this patch doesn't look correct to me.  Why is it appropriate to
> ignore evironment variables passed to the test script?

I was not intend to ignore the evironment variables. I just posted the problem
I've got and how did I avoid that.
The argument passed by Makefile is:
$(objpfx)tst-locale-locpath.out : tst-locale-locpath.sh $(objpfx)locale
        $(SHELL) $< '$(common-objpfx)' '$(test-wrapper)' '$(test-wrapper-env)'
But the argument taken by the script is:
common_objpfx=$1
test_wrapper_env=$2
run_program_env=$3
The test-wrapper is passed twice and no run_program_env is passed. I don't know
whether the test-wrapper script is unnecessary(last patch) or the argument passed
by Makefile is wrong:

diff --git a/locale/Makefile b/locale/Makefile
index 0ad99ec..d78cf9b 100644
--- a/locale/Makefile
+++ b/locale/Makefile
@@ -113,5 +113,5 @@ lib := locale-programs
 include $(patsubst %,$(..)libof-iterator.mk,$(cpp-srcs-left))
 
 $(objpfx)tst-locale-locpath.out : tst-locale-locpath.sh $(objpfx)locale
-       $(SHELL) $< '$(common-objpfx)' '$(test-wrapper)' '$(test-wrapper-env)' > $@; \
+       $(SHELL) $< '$(common-objpfx)' '$(test-wrapper-env)' '$(run-program-env)' > $@; \
        $(evaluate-test)

Thanks,
Mao Han


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]