Re: [sdcc-devel] SDCCsymt.c: Is it a bug?

[Maarten B. <sou...@ds...>]

Yes, Frieder also pointed that out to me in a PM. Sorry, I missed that.
I'll commit this fix in a few minutes.

> Hi Maarten, hi all,
>
>>>     /* if both are bitvars choose the larger one */
>>>   else if (IS_BITVAR (etype1) && IS_BITVAR (etype2))
>>>     rType = SPEC_BLEN (etype1) >= SPEC_BLEN (etype2) ?
>>>             copyLinkChain (type1) : copyLinkChain (type1);
>> This is no bug. If you had taken a look at the
>> definition of IS_BITVAR you could have seen that it also
>> includes bitfields.
>
> I guess the OP's problem was rather the double use of type1 in the last
> line (both alternatives result in copyLinkChain(type1)). Without having
> inspected the surrounding code, this also seems weird to me. From the
> comment I would have expected
> 	rType = SPEC_BLEN (etype1) >= SPEC_BLEN (etype2) ?
>         	copyLinkChain (type1) : copyLinkChain (type2);
> with type2 instead of type1 at the end.
>
> Regards,
> Raphael