Hi Bernd,
> this should fix the test failures in this test case.
>
> Is it OK for trunk?
unfortunately not: there are two bugs and a couple of nits:
* When testing with
runtest --tool gcc outputs.exp
I get
ERROR: tcl error sourcing
/vol/gcc/src/hg/master/local/gcc/testsuite/gcc.misc-tests/outputs.exp.
ERROR: can't unset "env(MAKEFLAGS)": no such element in array
while executing
"unset env($var)"
(procedure "unsetenv" line 3)
invoked from within
"unsetenv MAKEFLAGS"
(file
"/vol/gcc/src/hg/master/local/gcc/testsuite/gcc.misc-tests/outputs.exp" line 72)
invoked from within
"source /vol/gcc/src/hg/master/local/gcc/testsuite/gcc.misc-tests/outputs.exp"
("uplevel" body line 1)
invoked from within
"uplevel #0 source
/vol/gcc/src/hg/master/local/gcc/testsuite/gcc.misc-tests/outputs.exp"
invoked from within
"catch "uplevel #0 source $test_file_name""
The unsetenv needs to be wrapped in
if [info exists env(MAKEFLAGS)] {
to avoid this.
*
diff --git a/gcc/testsuite/gcc.misc-tests/outputs.exp
b/gcc/testsuite/gcc.misc-tests/outputs.exp
--- a/gcc/testsuite/gcc.misc-tests/outputs.exp
+++ b/gcc/testsuite/gcc.misc-tests/outputs.exp
@@ -67,6 +67,10 @@ if {[board_info $dest exists output_form
append link_options " additional_flags=-Wl,-oformat,[board_info $dest
output_format]"
}
+# Avoid possible influence from the make jobserver,
+# otherwise ltrans0.ltrans_args files may be missing.
+unsetenv MAKEFLAGS
The comment is misleading: it's not just *.ltrans_args, but also
*.ltrans.args.0. Maybe there's a collective term for those files in
lto-wrapper instead?
@@ -163,6 +167,9 @@ proc outest { test sources opts dirs out
if { $ogl != {} } {
pass "$test: $d$o"
file delete $ogl
+ } elseif { [string match "*.ld1_args" $o] } {
+ # This file may be missing if !HAVE_GNU_LD
+ pass "$test: $d$o"
Always PASSing the test even if it isn't run is wrong. Either wrap
the whole group of tests with response files in
if [check_effective_target_gld] {
or make the test for the *.ld1_args file conditional on that
(e.g. along the lines of $ltop used elsewhere). I'd welcome input
from Alexandre which is preferred.
A few nits on the patch submission:
* Please review https://gcc.gnu.org/contribute.html for the syntax of
subject lines: in the present case this should be something like
[PATCH] testsuite: Fix test failures from outputs.exp [PR98225]
* Both the mail and the patch description should contain a
self-contained description of the bug and the fix so potential
reviewers don't have to re-read a (potentially excessively long)
bugzilla report.
* Your ChangeLog entry isn't particularly helpful:
2021-01-07 Bernd Edlinger <bernd.edlin...@hotmail.de>
PR testsuite/98225
* gcc.misc-tests/outputs.exp: Fix test case.
This tells the reader almost nothing. Instead, it should list *what
changed* in the patch; for the current patch something like
* gcc.misc-tests/outputs.exp: Unset MAKEFLAGS.
Always pass *.ld1_args tests.
There's more than you ever wanted to know on ChangeLogs in the GNU
Coding Standards ;-)
Thanks for working on this.
Rainer
--
-----------------------------------------------------------------------------
Rainer Orth, Center for Biotechnology, Bielefeld University
