On 6/25/11 15:19, "Philipp Klaus Krause" <p...@spth.de> wrote: >sdcc has a --short-is-8bits option. It was broken for some time, but has >been fixed in rev. #6608. > >However IMO this looks like kind of option that we might want to remove >to me: >1) It enables nonstandard (as in not standard-conforming) behaviour >2) Everything to be gained by using this option (speed and code size) >can be gained in a standard way: Just use signed char or int_fast8_t or >int_least8_t instead of short int. >This requires changing the source, which is not an issue if it's >newly-written code. For existing code one would have to check everywhere >if it would work with 8-bit shorts as well, which is error-prone and not >less effort than changing the code to use 8-bit ints in a standard way. > >If there are any users of this option out there, please present your >use-case. > >If no one uses it I'd remove it.
I'm not using it, so I don't care, but to me this [removing it] looks a bit pointless, at least on the given grounds. It will brake somebody's code but will not help anyone else. If there is extra man power in the project to do something like this why not use it for something more productive like trying to optimize the code generation. Just my 2 snt. br Kusti ------------------------------------------------------------------------------ All the data continuously generated in your IT infrastructure contains a definitive record of customers, application performance, security threats, fraudulent activity and more. Splunk takes this data and makes sense of it. Business sense. IT sense. Common sense.. http://p.sf.net/sfu/splunk-d2d-c1 _______________________________________________ Sdcc-user mailing list Sdcc-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/sdcc-user
