RE: libgccjit: Clarifying the permitted type conversions

[Alex Coplan <Alex.Coplan@arm.com>]

Gentle ping.

> -----Original Message-----
> From: Jit <jit-bounces@gcc.gnu.org> On Behalf Of Alex Coplan
> Sent: 09 July 2020 15:15
> To: jit@gcc.gnu.org
> Cc: nd <nd@arm.com>
> Subject: libgccjit: Clarifying the permitted type conversions
> 
> Hello,
> 
> I have some questions about the type conversions allowed by the
> libgccjit interface.
> 
> Firstly, I believe it is the case that the intent is that all implicit
> conversions are forbidden. Is this correct?
> 
> Secondly, I wanted to clarify the situation with respect to explicit
> casts; that is, those conversions allowed by gcc_jit_context_new_cast().
> The docs [0] say:
> 
> Currently only a limited set of conversions are possible:
>  - int <-> float
>  - int <-> bool
>  - P* <-> Q*, for pointer types P and Q
> 
> However, empirically (at least on aarch64), libgccjit appears to allow
> me to compile casts between any pair of types in the following set
> without any complaint:
> 
> {
>   SIGNED_CHAR,
>   UNSIGNED_CHAR,
>   SHORT,
>   UNSIGNED_SHORT,
>   INT,
>   UNSIGNED_INT,
>   LONG,
>   UNSIGNED_LONG,
>   LONG_LONG,
>   UNSIGNED_LONG_LONG
> }
> 
> Is this intended behaviour? If so, does the documentation need to be
> updated here? If not, perhaps there are some checks missing in the
> libgccjit interface?
> 
> Many thanks,
> Alex
> 
> [0] : https://gcc.gnu.org/onlinedocs/jit/topics/expressions.html