Re: [PATCH] fort_dyn_array: enable dynamic array of types

[Joel Brobecker <brobecker@adacore.com>]

Keven,

On Thu, Nov 05, 2015 at 04:56:31PM +0100, Keven Boell wrote:
> This patch enables dynamic arrays of types in Fortran.
> 
> 2015-03-20  Keven Boell  <keven.boell@intel.com>
> 
> 	* gdbtypes.c (resolve_dyn_properties): New.
> 	(resolve_dynamic_range): Add call to resolve_dyn_properties
> 	to resolve data_location.
> 	(resolve_dynamic_array): Add call to resolve_dyn_properties
> 	to resolve data_location.
> 	(resolve_dynamic_union): Add call to resolve_dyn_properties
> 	to resolve_data_location.
> 	(resolve_dynamic_struct): Add call to resolve_dyn_properties
> 	to resolve_data_location.  Adjust address of inner types of
> 	a struct to reflect the offset of the inner's type data_location.
> 	(resolve_dynamic_type_internal): Remove data_location computation.
> 	* value.c (set_value_component_location): Adjust the value address
> 	for single value prints.
> 	(value_primitive_field): Use lazy fetch to re-evaluate a field to
> 	get the actual value.
> 	(value_fetch_lazy): Use address provided by DWARF data_location
> 	attribute if present.
> 
> testsuite/gdb.fortran:
> 
> 	* vla-type.exp: New file.
> 	* vla-type.f90: New file.

This is really not the way I tried to explained how these changes
should be submitted. I don't know about others, but I can't understand
the reasons for the changes if you don't show me what's happening
today, the debugging info, why the debugger is failing, and explain
how you are fixing the issue.

For instance, I suspect that this patch can be split into smaller,
independent patches, which themselves fixes a smaller set of issues.
And I'd like to make sure we have a test for each hunk that you're
adding.

-- 
Joel