Re: [PATCH 1/2] gdb/testsuite: allowed for function_range to deal with mangled functions

[Bruno Larsen <blarsen@redhat.com>]

On 31/10/2022 15:46, Andrew Burgess wrote:
> Bruno Larsen via Gdb-patches <gdb-patches@sourceware.org> writes:
>
>> When calling get_func_info inside a test case, it would cause failures
>> if the function was printed using a C++ style mangled name. The current
>> patch fixes this by allowing for mangled names along with the current
>> rules.
> Instead of this approach, I would like to propose an alternative which I
> believe also addresses this issue.  What do you think of this?

This sounds like a much better approach. Everyday I'm learning about 
some new thing in GDB.

Will send it like this on v2

Cheers,
Bruno

> Thanks,
> Andrew
>
> ---
>
> commit b6d59ad650cfc552487c82a4eef49442b0642bca
> Author: Bruno Larsen via Gdb-patches <gdb-patches@sourceware.org>
> Date:   Thu Oct 13 18:01:15 2022 +0200
>
>      gdb/testsuite: allowed for function_range to deal with mangled functions
>      
>      When calling get_func_info inside a test case, it would cause failures
>      if the function was printed using a C++ style mangled name.  Fix this
>      by asking the disassembler to demangle symbols.
>
> diff --git a/gdb/testsuite/lib/dwarf.exp b/gdb/testsuite/lib/dwarf.exp
> index 593f4ea72f1..17faf06992e 100644
> --- a/gdb/testsuite/lib/dwarf.exp
> +++ b/gdb/testsuite/lib/dwarf.exp
> @@ -400,7 +400,7 @@ proc function_range { func src {options {debug}} } {
>       if { $func_length != 0 } {
>   	set func_pattern "$func_pattern\\+$func_length"
>       }
> -    set test "x/2i $func+$func_length"
> +    set test "with print asm-demangle on -- x/2i $func+$func_length"
>       gdb_test_multiple $test $test {
>   	-re ".*($hex) <$func_pattern>:\[^\r\n\]+\r\n\[ \]+($hex).*\.\r\n$gdb_prompt $" {
>   	    set start $expect_out(1,string)
>