[PATCH] mark libstdc++ tests unsupported if they fail with "relocation truncated"

[PATCH] mark libstdc++ tests unsupported if they fail with "relocation truncated"

[Szabolcs Nagy]

[-- Attachment #1: Type: text/plain, Size: 715 bytes --]

Dejagnu tweak:
Check the compiler output in libstdc++-dg-test using
${tool}_check_unsupported_p and mark the output
unsupported accordingly to avoid "relocation truncated"
failures cluttering the test results on aarch64-none-elf
with -mcmodel=tiny. (gcc torture tests already use this
predicate for this reason.)

This is a revision of
https://gcc.gnu.org/ml/libstdc++/2015-01/msg00198.html
which had problems because the unsupported result was
propagated to and handled by dg-runtest differently
than expected.

Is it OK?

libstdc++-v3/Changelog:

2015-10-13  Szabolcs Nagy  <szabolcs.nagy@arm.com>

	* testsuite/lib/libstdc++.exp (libstdc++-dg-test): Check for
	unsupported compiler output.

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: reloc_trunc_dg.diff --]
[-- Type: text/x-patch; name=reloc_trunc_dg.diff, Size: 704 bytes --]

diff --git a/libstdc++-v3/testsuite/lib/libstdc++.exp b/libstdc++-v3/testsuite/lib/libstdc++.exp
index 88738b7..769ac94 100644
--- a/libstdc++-v3/testsuite/lib/libstdc++.exp
+++ b/libstdc++-v3/testsuite/lib/libstdc++.exp
@@ -409,6 +409,11 @@ proc libstdc++-dg-test { prog do_what extra_tool_flags } {
     # and-target file), but the defaults are lacking in goodness.
     set comp_output [$select_compile "$prog" "$output_file" "$compile_type" $options];
 
+    set unsupported_message [libstdc++_check_unsupported_p $comp_output]
+    if { $unsupported_message != "" } {
+	set comp_output "::unsupported::$unsupported_message"
+    }
+
     return [list $comp_output $output_file]
 }
 

Re: [PATCH] mark libstdc++ tests unsupported if they fail with "relocation truncated"

[Jonathan Wakely]

On 13/10/15 10:12 +0100, Szabolcs Nagy wrote:
>Dejagnu tweak:
>Check the compiler output in libstdc++-dg-test using
>${tool}_check_unsupported_p and mark the output
>unsupported accordingly to avoid "relocation truncated"
>failures cluttering the test results on aarch64-none-elf
>with -mcmodel=tiny. (gcc torture tests already use this
>predicate for this reason.)
>
>This is a revision of
>https://gcc.gnu.org/ml/libstdc++/2015-01/msg00198.html
>which had problems because the unsupported result was
>propagated to and handled by dg-runtest differently
>than expected.
>
>Is it OK?

OK, assuming this was tested on a primary platform and didn't change
the libstdc++ testresults.