[antlr-interest] C++ compliance of antlr generated code

Ric Klaren ric.klaren at gmail.com
Mon Jan 1 08:46:28 PST 2007


On 12/31/06, Maurizio de Pascale <mdepascale at dii.unisi.it> wrote:
> So I was not saying that this is a bug, or that it will break some code
> in the future, but since I suppose that the C++ target for ANTLR3 is
> already being rewritten, I still suggest that it could be worth to do a
> bit of find-and-replace to handle this.

There's variable names coming from some parts of the 'base'
codegenerator. If these do not need additionial processing then I'll
probably will not touch them (so '_' may creep in there). In the
support code itself you will probably never see '_' at the start of a
variable since I dislike it.

> Why is _ttype better than ttype, tokenType, token_type, or whatever one
> can choose?

Whichever works ;) as long as it's used structurally across the
library. I like consistency.

> P.S. is ANTLR3 C++ target complete?

I was catching up before I bought my house, but the C++ target never
left prototype stage. At the moment work will probably restart end of
januari after I've done most of the decorating moving. I can give no
timeframe since work interferes a lot as well.



More information about the antlr-interest mailing list