[PATCH v2] gdb: CTF support

[Wei-min Pan]

On 10/3/2019 5:40 AM, Simon Marchi wrote:
> On 2019-10-02 7:09 p.m., Wei-min Pan wrote:
>> It doesn't seem using gdb::unique_xmalloc_pointer is appropriate in
>> these cases
>> because we want to keep these symbol/type names around in the symbol table.
> I mean, instead of calling `free` by hand, that copy could be managed with
> a gdb::unique_xmalloc_pointer.  The name duplicated into the symbol table
> won't change.

But the `free` call is needed to free up space allocated by libctf's 
ctf_type_aname_raw.

> Simon
>