[PATCH] Propagate record_print_flags

[Simon Marchi]

On 2018-03-01 09:30, Metzger, Markus T wrote:
> Hello Simon,
> 
>> These flags are returned as an int by get_call_history_modifiers, and
>> get cast back to record_print_flags in the btrace code.  Instead, we 
>> can
>> make the arguments of that type from start to end.
>> 
>> gdb/ChangeLog:
>> 
>> 	* record.c (get_call_history_modifiers): Return a
>> 	record_print_flags.
>> 	(cmd_record_call_history): Adjust.
>> 	* record-btrace.c (record_btrace_call_history): Adjust.
>> 	(record_btrace_call_history_range): Adjust.
>> 	(record_btrace_call_history_from): Adjust.
>> 	* target-debug.h (target_debug_print_record_print_flags): New.
>> 	* target-delegates.c: Re-generate.
>> 	* target.c (target_call_history): Change flags type.
>> 	(target_call_history_from): Likewise.
>> 	(target_call_history_range): Likewise.
>> 	* target.h (struct target_ops) <target_call_history>: Likewise.
>> 	(target_call_history_from): Likewise.
>> 	(target_call_history_range): Likewise.
> 
> Looks good to me.

Thanks for checking.

>> @@ -172,6 +172,8 @@
>>    target_debug_do_print (host_address_to_string (X.data ()))
>>  #define target_debug_print_inferior_p(inf) \
>>    target_debug_do_print (host_address_to_string (inf))
>> +#define target_debug_print_record_print_flags(X) \
>> +  target_debug_do_print (plongest (X))
> 
> Should we print them as hex?

It would be even better to print them as flags (VALUE_A | VALUE_B).  
I'll push this patch as-is and suggest something that will cover this 
and the other target_debug_print that deal with flags.

Simon