[PATCH] Fix 30808 gprofng tests failed

Vladimir Mezentsev vladimir.mezentsev@oracle.com
Thu Jan 18 00:49:28 GMT 2024


On 1/17/24 03:24, Xi Ruoyao wrote:
> On Fri, 2023-09-01 at 13:04 -0700, Vladimir Mezentsev via Binutils wrote:
>
>> diff --git a/gprofng/testsuite/config/default.exp b/gprofng/testsuite/config/default.exp
>> index 1c862d54592..c4a6588bb64 100644
>> --- a/gprofng/testsuite/config/default.exp
>> +++ b/gprofng/testsuite/config/default.exp
>> @@ -32,17 +32,25 @@ if {![info exists CFLAGS]} {
>>       set CFLAGS "-g -O2"
>>   }
>>   
>> +remote_exec host "sh -c \"rm -rf tmpdir; mkdir -p tmpdir\""
>>   if { "$CHECK_TARGET" == "check-install" } {
>>       set ::env(GPROFNG) "$PREFIX/bin/gprofng"
>>   } else {
>> -    set ::env(GPROFNG) "$BUILDDIR/src/gprofng"
>> +    # Make a temporary install dir to run gprofng from, and point at it
>> +    remote_exec host "sh -c \"$MAKE -C .. install-gprofng MAKEFLAGS=program_transform_name= DESTDIR=$BUILDDIR/tmpdir/root\""
> Unfortunately this has bring back PR29107.  I've created PR31252.
>

I cannot reproduce the problem on the master branch ( OL8 / x86_64 ).
I did:

mkdir bld_x86_64
cd bld_x86_64
../configure --prefix=/usr --with-system-zlib --enable-shared 
--disable-{gdb,gdbserver,libbacktrace,libdecnumber,readline,sim}
make -j16 -O tooldir=/usr V=1 2>&1 | tee make.log
make -k -j16 -O check 2>&1 | tee make_check.log

I see errors but they are not related to 30808.
% grep FAIL make_check.log
FAIL: 
/ws/gprofng-dev-backedup/vmezents/GPROF/binutils-gdb/libctf/testsuite/libctf-regression/libctf-repeat-cu.exp
FAIL: 
/ws/gprofng-dev-backedup/vmezents/GPROF/binutils-gdb/libctf/testsuite/libctf-regression/nonstatic-var-section-ld-r.c
FAIL: replacing SOURCE_DATE_EPOCH deterministic member (wrong size, 
expected: 1120)
FAIL: archive with empty element
FAIL: cxxfilt: demangling _Z1fIvJiELb0EEvPDOT1_EFT_DpT0_E
FAIL: cxxfilt: demangling 
_Z14int_if_addableI1YERiP1AIXszpldecvPT_Li0EdecvS4_Li0EEE

Could you send me the log files with errors.

Thank you,
-Vladimir



More information about the Binutils mailing list