Re: [PATCH v5 3/3] [gdb/exp] Handle recursive namespace import

[Tom de Vries <tdevries@suse.de>]

On 5/1/26 9:37 PM, Tom Tromey wrote:
>>>>>> "Tom" == Tom de Vries <tdevries@suse.de> writes:
> 
> Tom> Fix this by additionally iterating over the super blocks instead of only using
> Tom> the current block.
> 
> FWIW I haven't really commented much on the core part of this series
> because I find this code pretty confusing and I'm not really sure I
> understand it well enough to do a good job reviewing it.
> 
> Tom> +	      for (const struct block *b = block; b != nullptr;
> Tom> +		   b = b->superblock ())
> Tom> +		cp_lookup_symbol_via_imports (current->import_src, name,
> Tom> +					      b, domain, 1, 0, 0,
> 
> In this spot, suppose 'block' is from an inlined function.  Then when
> iterating, I think we'll end up seeing blocks from the enclosing
> function (the one that was inlined-into).  However, that could yield the
> wrong result, because the inlined and inlined-into functions could come
> from completely different namespace contexts with different 'using'
> directives.

Thanks for pointing this out, fixed in a v6 ( 
https://sourceware.org/pipermail/gdb-patches/2026-May/227330.html ).

Thanks,
- Tom