Jonathan M Davis wrote:
On Thursday 21 October 2010 15:01:21 Walter Bright wrote:
5. generally follow along with the C++ one so that they can be maintained
in tandem
Does this mean that you want a pseudo-port of the C++ front end's lexer to D for
this? Or are you looking for just certain pieces of it to be similar?
Yes, but not a straight port. The C++ version has things in it that are
unnecessary for the D version, like the external string table (should use an
associative array instead), the support for lookahead can be put in the parser,
doesn't tokenize comments, etc.
Essentially I'd like the D lexer to be self-contained in one file.
I haven't looked at the front end code yet, so I don't know how it works there,
but I wouldn't expect it to uses ranges, for instance, so I would expect that
the basic design would naturally stray a bit from whatever was done in C++
simply by doing things in fairly idiomatic D. And if I do look at the front end
to see how that's done, there's the issue of the license. As I understand it,
the front end is LGPL, and Phobos is generally Boost, which would mean that I
would be looking at LGPL-licensed code when designing Boost-licensed, even
though it wouldn't really be copying the code per se since it's a change of
language (though if you did the whole front end, obviously the license issue can
be waved quite easily).
Since the license is mine, I can change the D version to the Boost license, no
problem.
License issues aside, however, I do think that it would make sense for
std.lang.d.lex to do things similiarly to the C++ front end, even if there are a
number of basic differences.
Yup. The idea is the D version lexes exactly the same grammar as the dmd one.
The easiest way to ensure that is to do equivalent logic.