Re: [PATCH 09/11] gdb/testsuite: don't test gdb.base/msym-bp-shl with clang

[Bruno Larsen <blarsen@redhat.com>]

On 3/2/22 16:33, Pedro Alves wrote:
> On 2022-01-26 19:50, Bruno Larsen via Gdb-patches wrote:
>> Clang will optimize away the static function in one of the files, and
>> the test is here to specifically test GDB's behavior when that function
>> is present, so it makes no sense to have this test run with that
>> compiler.
> 
> Please expand this info to include a snippet of the failing test.
> Also, it's best if the the body of the commit log, doesn't assume the subject is read
> as part of it.  You will notice that if you read the body in isolation, without the
> subject line, and also without the diff context, the commit log is vague as is.

Oh, ok. Will do!

> 
>> ---
>>   gdb/testsuite/gdb.base/msym-bp-shl.exp | 7 +++++++
>>   1 file changed, 7 insertions(+)
>>
>> diff --git a/gdb/testsuite/gdb.base/msym-bp-shl.exp b/gdb/testsuite/gdb.base/msym-bp-shl.exp
>> index 42adcb191dd..95d5c393505 100644
>> --- a/gdb/testsuite/gdb.base/msym-bp-shl.exp
>> +++ b/gdb/testsuite/gdb.base/msym-bp-shl.exp
>> @@ -22,6 +22,13 @@ if {[skip_shlib_tests]} {
>>       return 0
>>   }
>>   
>> +# clang will optimize away the static foo, making a single breakpoint
>> +# so there is no point testing it here.
> 
> Uppercase Clang.
> 
> The patch doesn't give enough info to understand why this is the best approach.
> I could imagine using atttribute used to make sure the function isn't optimized
> away, as alternative approach, for example.  Please make the case for the approach
> taken.

Doh! I forgot that attribute((used)) worked for functions as well, I'll do that instead, since it also works.

> 
>> +if {[test_compiler_info {clang-*-*}]} {
>> +    untested "clang only compiles one foo"
>> +    return
>> +}
>> +
>>   standard_testfile msym-bp-shl-main.c msym-bp-shl-main-2.c msym-bp-shl-lib.c
>>   set srcfile ${srcdir}/${subdir}/${srcfile}
>>   set srcfile2 ${srcdir}/${subdir}/${srcfile2}
> 


-- 
Cheers!
Bruno Larsen