On Sun, Dec 15, 2019 at 08:26:44PM +0100, Luc Van Oostenryck wrote: > On Sun, Dec 15, 2019 at 07:20:34PM +0000, Ramsay Jones wrote: > > On 15/12/2019 11:04, Luc Van Oostenryck wrote: > > > > Hmm, it was not immediately clear that the '!base' path did not > > introduce an (effective) functional change. I suspect that it > > does not, but I wasn't sure if examine_node_type(sym) for the > > above 'bad_ctype' symbol would add alignment, bit_size or rank > > to the symbol (and even if it did, would it matter?). > > Mmmm, yes. It shouldn't matter but I prefer to avoid this. OTOH, examine_symbol_type() can anyway be called on it later. So, I'll leave it so as I find it more readable. -- Luc
