Front page | perl.bootstrap |
Postings from July 2000
Dan Sugalski wrote: > At 09:51 AM 7/22/00 -0400, Bradley M. Kuhn wrote: > >At the conference, Dan suggested that it's imperative that this subset of > >Perl be source-to-source translatable into C. Even pa GCC front-end wouldn't > >work, because we want it to be compatible with other compilers. > > I wasn't adamant that it went to C, so much as it went completely to some > language (rather than language implementation) we could be reasonably sure > was on most people's machines, which sort of means either C or C++. My apologies for misquoting you, Dan. Thanks for clearing that up. Dan and I were talking over food at a party, so it's easy to get things wrong in that situation, especially when one is hungry. ;) Of course, C is somewhat of a lingua franca of modern computer systems, so C might turn out to be the right choice for a "source-to-source" translation in this case. Do folks think it's useful to hack up a proof of concept of a parser for something Perl-ish written in a subset of something Perl-ish? Or, is the VHDL system Nick mentioned elsewhere in the thread already enough of a proof of concept? (BTW, I would like to note that, IMO, any code that gets written before the working groups (which Larry's wife has suggested we rename "cloysters" :) get started should definitely be marked as "disposable". However, it might be useful to write some code as proof of concepts at this point, which is what I am suggesting above, if others think it's a good idea.) Finally, I agree with Nick that writing the back-end in a subset of Perl is probably a bad idea. I think we should give serious thought to the front end in a subset of Perl, though. -- Bradley M. Kuhn - http://www.ebb.org/bkuhn