On Sun, 06 Jan 2019 17:01:20 -0500 "Eric S. Raymond" <e...@thyrsus.com> wrote:
>
> A simple, possibly correct LR parser for C11
>
> http://gallium.inria.fr/~fpottier/publis/jourdan-fpottier-2016.pdf
This paper says its lexer+parser is about 1000 lines, which
doesn't include the preprocessor. For comparison, subc (a
compiler for a self compiling subset of C) is under 5K lines,
including a preprocessor. It is a recursive descent compiler
which may be easier to grok. Best of all, there is an
associated book called "practical compiler construction" (for
a slightly older compiler). There is even a Go version of
subc! subc is completely public domain.
https://www.t3x.org/subc/index.html
May be worth checking subc out (though it will have to be
extended to cover missing features such as goto, typedef...)
The other suggestion I have is to figure out how to map
individual C constructs to Go and not try mapping idiomatic C
code to idiomatic Go code. The former is a hard enough
problem as it is. Then may be you can convert robotically
translated code to idiomatic code (sort of what the Go folks
must've done/are doing). This way you can internalize most of
the translation scheme before writing a single line of
production code. And it is much easier to change the xlation
scheme while you have just a paper design! [I once spent a
bunch of tine on writing a Beta language to C translator and
followed this path]
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
