Le 29 juil. 2013 à 14:25, Valentin Tolmer <[email protected]> a écrit :

> The current structure of the symbol aliases is puzzling: why have the 
> information duplicated only to resolve it later on?

This is because two symbols may be discovered to be aliases
late I guess.  Or because the support for aliases was added
without careful thinking.

> I'm working on a patch to address this issue: the symbol and the alias will 
> share a pointer to a separate structure containing the info. The symbol 
> itself will have only its intrinsic info such as the name or location.
> 
> You can have a look at the patch here: 
> https://github.com/nitnelave/bison/tree/symbol
> I have yet to add some tests for what I added, but it is near completion

Great. Thanks!


Reply via email to