On 8/23/23 17:43, Pedro Alves wrote:
> On 22/08/23 18:34, Tom de Vries wrote:
> 
>> I've given the gdb::array_view<const gdb_byte> a try, and that works fine, and greatly simplifies the patch, so thanks for that suggestion.
> 
> Great!
> 
>> The first patch is no longer necessary, but we could consider committing it nevertheless, since the work is done and may be needed in the future.
> 
> I agree.
> 
>>   
>> -gdb::byte_vector
>> +gdb::array_view<const gdb_byte>
>>   remote_target::thread_info_to_thread_handle (struct thread_info *tp)
>>   {
>>     remote_thread_info *priv = get_remote_thread_info (tp);
>> diff --git a/gdb/target-debug.h b/gdb/target-debug.h
>> index acb01d47e7c..31c25e9aefd 100644
>> --- a/gdb/target-debug.h
>> +++ b/gdb/target-debug.h
>> @@ -236,4 +236,18 @@ target_debug_print_gdb_byte_vector_r (gdb::byte_vector &vector)
>>   {
>>     target_debug_print_const_gdb_byte_vector_r (vector);
>>   }
>> +
>> +static void
>> +target_debug_print_gdb_array_view_const_gdb_byte (gdb::array_view<const gdb_byte> &vector)
> 
> array_views should be passed by value normally (because they're small).  Did you run into something
> that forced this to be a reference?  Or was it just copied from the vector case?  If it
> works, we should just drop the &.
> 

Done.

> Also, one of the points of array_view is that you can pass a vector as array_view
> argument, like you can pass std::string as string_view argument -- would it work to make
> target_debug_print_const_gdb_byte_vector_r call this new
> target_debug_print_gdb_array_view_const_gdb_byte instead of duplicating that code?
> 

Done, updated version attached.

Thanks,
- Tom