Re: [PATCH V4 5/6] Resolve dynamic target types of pointers.

[Bernhard Heckel <bernhard.heckel@intel.com>]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset="utf-8"; format="flowed", Size: 6060 bytes --]

On 13/10/2016 16:07, Yao Qi wrote:
> Bernhard Heckel <bernhard.heckel@intel.com> writes:
>
>> diff --git a/gdb/c-valprint.c b/gdb/c-valprint.c
>> index 2cb418d..120163f 100644
>> --- a/gdb/c-valprint.c
>> +++ b/gdb/c-valprint.c
>> @@ -645,6 +645,28 @@ c_value_print (struct value *val, struct ui_file *stream,
>>         else
>>   	{
>>   	  /* normal case */
>> +	  if (TYPE_CODE (type) == TYPE_CODE_PTR
>> +	      && 1 == is_dynamic_type (type))
> Use is_dynamic_type return value as bool.
>
>> +	    {
>> +	      CORE_ADDR addr;
>> +	      if (NULL != TYPE_DATA_LOCATION (TYPE_TARGET_TYPE (type)))
>> +		addr = value_address (val);
>> +	      else
>> +		addr = value_as_address (val);
>> +
>> +	      /* We resolve the target-type only when the
>> +	         pointer is associated.  */
>> +	      if ((addr != 0)
>> +		  && (0 == type_not_associated (type)))
> !type_not_associated (type)
>
>> +
>> +
>> +gdb_breakpoint [gdb_get_line_number "Before pointer assignment"]
>> +gdb_continue_to_breakpoint "Before pointer assignment"
>> +gdb_test "print logp" "= \\(PTR TO -> \\( $logical \\)\\) 0x0" "print logp, not associated"
>> +gdb_test "print *logp" "Cannot access memory at address 0x0" "print *logp, not associated"
>> +gdb_test "print comp" "= \\(PTR TO -> \\( $complex \\)\\) 0x0" "print comp, not associated"
>> +gdb_test "print *comp" "Cannot access memory at address 0x0" "print *comp, not associated"
>> +gdb_test "print charp" "= \\(PTR TO -> \\( character\\*1 \\)\\) 0x0" "print charp, not associated"
>> +gdb_test "print *charp" "Cannot access memory at address 0x0" "print *charp, not associated"
>> +gdb_test "print charap" "= \\(PTR TO -> \\( character\\*3 \\)\\) 0x0" "print charap, not associated"
>> +gdb_test "print *charap" "Cannot access memory at address 0x0" "print *charap, not associated"
>> +gdb_test "print intp" "= \\(PTR TO -> \\( $int \\)\\) 0x0" "print intp, not associated"
>> +gdb_test "print *intp" "Cannot access memory at address 0x0" "print
>> *intp, not associated"
> These lines are too long.
>
>> +set test "print intap, not associated"
>> +gdb_test_multiple "print intap" $test {
>> +  -re " = \\(PTR TO -> \\( $int \\(:,:\\)\\)\\) <not associated>\r\n$gdb_prompt $" {
>> +    pass $test
>> +  }
>> +  -re " = <not associated>\r\n$gdb_prompt $" {
>> +    pass $test
>> +  }
>> +}
> Why do we need to different regex here?  Need to handle the differences
> of different compilers?  IMO, the former looks reasonable to me.
A pointer to an array (especially to Variable Length Array, VLA) is 
described in Gfortran's  compiled debug info
like normal array types + associated property of an pointer.
Not sure if this is a bug in Gfortran or not? I was also thinking to 
make the
output equal, and check if an array has pointer attributes but ..... I 
don't really like this approach. This would
cause a lot of "if .. else" construct in the code for example: 
De-reference this pointer, Address operator, ...

>> +gdb_test "print realp" "= \\(PTR TO -> \\( $real \\)\\) 0x0" "print realp, not associated"
>> +gdb_test "print *realp" "Cannot access memory at address 0x0" "print *realp, not associated"
> They are too long.
>
>> +gdb_test "print \$my_var = intp" "= \\(PTR TO -> \\( $int \\)\\) 0x0"
>> +
>> +
>> +gdb_breakpoint [gdb_get_line_number "Before value assignment"]
>> +gdb_continue_to_breakpoint "Before value assignment"
>> +gdb_test "print *(twop)%ivla2" "= <not allocated>"
>> +
>> +
>> +gdb_breakpoint [gdb_get_line_number "After value assignment"]
>> +gdb_continue_to_breakpoint "After value assignment"
>> +gdb_test "print logp" "= \\(PTR TO -> \\( $logical \\)\\) $hex\( <.*>\)?"
>> +gdb_test "print *logp" "= \\.TRUE\\."
>> +gdb_test "print comp" "= \\(PTR TO -> \\( $complex \\)\\) $hex\( <.*>\)?"
>> +gdb_test "print *comp" "= \\(1,2\\)"
>> +gdb_test "print charp" "= \\(PTR TO -> \\( character\\*1 \\)\\) $hex\( <.*>\)?"
>> +gdb_test "print *charp" "= 'a'"
>> +gdb_test "print charap" "= \\(PTR TO -> \\( character\\*3 \\)\\) $hex\( <.*>\)?"
>> +gdb_test "print *charap" "= 'abc'"
>> +gdb_test "print intp" "= \\(PTR TO -> \\( $int \\)\\) $hex\( <.*>\)?"
>> +gdb_test "print *intp" "= 10"
>> +set test_name "print intap, associated"
>> +gdb_test_multiple "print intap" $test_name {
>> +  -re "= \\(\\( 1, 1, 3(, 1){7}\\) \\( 1(, 1){9}\\) \\)\r\n$gdb_prompt $" {
>> +    pass $test_name
>> +  }
>> +  -re "= \\(PTR TO -> \\( $int \\(10,2\\)\\)\\) $hex\( <.*>\)?\r\n$gdb_prompt $" {
>> +    gdb_test "print *intap" "= \\(\\( 1, 1, 3(, 1){7}\\) \\( 1(, 1){9}\\) \\)"
>> +    pass $test_name
>> +  }
>> +}
> Again, different outputs due to different compilers?  The latter looks
> reasonable to me.
>
>> +set test_name "print intvlap, associated"
>> +gdb_test_multiple "print intvlap" $test_name {
>> +  -re "= \\(2, 2, 2, 4(, 2){6}\\)\r\n$gdb_prompt $" {
>> +    pass $test_name
>> +  }
>> +  -re "= \\(PTR TO -> \\( $int \\(10\\)\\)\\) $hex\( <.*>\)?\r\n$gdb_prompt $" {
>> +    gdb_test "print *intvlap" "= \\(2, 2, 2, 4(, 2){6}\\)"
>> +    pass $test_name
>> +  }
>> +}
> Likewise.
>
>> diff --git a/gdb/valprint.c b/gdb/valprint.c
>> index 6896da2..f592278 100644
>> --- a/gdb/valprint.c
>> +++ b/gdb/valprint.c
>> @@ -1141,12 +1141,6 @@ value_check_printable (struct value *val, struct ui_file *stream,
>>         return 0;
>>       }
>>   
>> -  if (type_not_associated (value_type (val)))
>> -    {
>> -      val_print_not_associated (stream);
>> -      return 0;
>> -    }
>> -
> Could you explain why do you remove this?
I we keep this, we shortcut any additional printing of the target type 
of a pointer in case it is not associated.
>
>>     if (type_not_allocated (value_type (val)))
>>       {
>>         val_print_not_allocated (stream);

Intel Deutschland GmbH
Registered Address: Am Campeon 10-12, 85579 Neubiberg, Germany
Tel: +49 89 99 8853-0, www.intel.de
Managing Directors: Christin Eisenschmid, Christian Lamprechter
Chairperson of the Supervisory Board: Nicole Lau
Registered Office: Munich
Commercial Register: Amtsgericht Muenchen HRB 186928
\x16º&Öéj×!zÊÞ¶êç×}ýëyb²Ö«r\x18\x1dn–­r\x17¬