Re: [PATCH] gdb: fix path expression for anonymous union members

[Keith Seitz <keiths@redhat.com>]

Hi,

Thanks for the patch.

On 10/7/26 1:27 AM, Stephan Rohr wrote:
> 
> Extract the required logic from cplus_describe_child and use it to get
> the correct index of the field within the parent type.  With this change
> the path expression is computed correctly:
> 
>    interpret mi "-var-info-path-expression var1.public.1_anonymous.public.c"
>    ^done,path_expr="((s).c)"
> 

I think this is the correct approach.

> I modified the related test to include both base type and access layers.
> However, to correctly parse the base types within the MI var tree I had to
> introduce a heuristic that if the child's type is equal to child's name
> then it is a base type, and the path expression is different in this case.

Let's look at this a bit in the actual change:

> diff --git a/gdb/testsuite/lib/mi-support.exp b/gdb/testsuite/lib/mi-support.exp
> index cd7e0a61189..465cf3fd69c 100644
> --- a/gdb/testsuite/lib/mi-support.exp
> +++ b/gdb/testsuite/lib/mi-support.exp
> @@ -2401,6 +2401,10 @@ namespace eval ::varobj_tree {
>         set is_compound 0
>       }
>   
> +    if {$type eq $name} {
> +	# This is the base type.
> +	return "((class $type) $path_parent(path_expr))"
> +    }
>       if {[string index $parent(type) end] == "\]"} {
>         # Parent is an array.
>         return "($path_parent(path_expr))\[$name\]"

IIUC, if we have a construct like:

class Outer {
public:
   struct Inner { int x; };
   Inner Inner;
};

Outer o;

Then the proposed heuristic change would return "((class Inner) o)".
However, the real path should be "((o).Inner)".

While this is probably unlikely to actually affect anything,
perhaps we can tighten this a little bit by applying the heuristic
only if its parent is not one of these artificial access parents?

That is, something like:

proc is_cplus_fake_child {variable} {
   upvar #0 $variable varobj
   return [string equal $varobj(type) ""]
}

# Note that proc get_path_expr walks parent_variable -- we need to save
# it before that.
if {$type eq $name && is_cplus_fake_child $saved_parent} {
    return #...
}

That might be a simple change to mitigate some future frustration.

Keith