On Fri, Aug 06, 2021 at 04:23:27PM +0100, Philip Herron wrote:
> On 05/08/2021 20:37, Mark Wielaard wrote:
> > Hi,
> >
> > On Fri, Jul 30, 2021 at 03:03:13AM +0200, Mark Wielaard wrote:
> >> That variant is attached and can also be found here:
> >> https://code.wildebeest.org/git/user/mjw/gccrs/commit/?h=pass-pratt-parse-loc
> >> The original is also here:
> >> https://code.wildebeest.org/git/user/mjw/gccrs/commit/?h=pass-pratt-parse-token
> >>
> >> Hopefully one of the two is acceptable. If not please let me know how
> >> to rewrite it in a cheaper more idiomatic way.
> > Any feedback on these patches? I really like to get the locations
> > correct for these pratt parsed expressions. If both variants are
> > problematic please let me know what I can do to improve them.
> >
> > Thanks,
> >
> > Mark
> >
> Hi Mark,
> 
> Sorry for lack of reply on this. After looking into this and also
> getting Arthur Cohen to review, I think:
> 
>   * Token passing: If we end up using the token for more than the
>     location this might be useful but at present I can't think of a
>     use-case for the token. If we do need it maybe we could look into
>     passing it by reference.
>   * Location passing: I think this is the best solution and is directly
>     associated with the issue you have fixed. The function says that you
>     can call me but tell me the location as well so its up to the caller
>     to handle this. We should add a comment to the function prototypes
>     to say what the location is but that's it.
> 
> What do you think?

OK, lets go with the location passing. I added a comment to the
function prototypes in rust-parse.h about the passed in
pratt_parsed_token.

https://code.wildebeest.org/git/user/mjw/gccrs/commit/?h=pass-pratt-parse-token

Cheers,

Mark