[antlr-interest] Re: More on ANTLR 2.7.2, VC++ 6.0 and portability

[iank at bearcave.com]

  I too have been making the transition to Antlr 2.7.2.  I've never
  been a big fan of Microsoft DLLs, so I've created a static link
  (e.g., a .lib).  I'd do this even if I could build the DLL.  Nor do
  I use the IDE project builder for building something like this.
  Note that Microsoft most certainly does not build large software
  components this way either.  They use and I use nmake, which is
  Microsoft's mutated version of UNIX make.

  I'm in the process of rewriting my ANTLR tutorial web pages, so I
  have not published the new nmake file yet.  If you need it before I
  get to it, send me a note.  I picked up a few additional changes to
  reduce the number of warnings and to allow the software to build
  from the ANTLR yahoo groups archive.  This was a change to one file
  (antlr/config.h?)  I can send you the modified source files as well.

  The other surprising experience I had is that a simple test grammar
  (e.g., my expression grammar) and its associated "main" no longer
  ran properly.  Some of this had to do with the way the AST factory
  is initialized (I picked this up from the examples).  However there
  seem to be grammar based problems as well (no grammar errors or
  warnings, but the grammer works differently now).  So the transition
  has not been as painless as I had hoped.  My fear now is that a
  large Java language grammar/AST generation front end I have will not
  port painlessly either.  Since this weighs in at about 5K lines of
  comments, grammar and C++ code this is sort of scary.

  Perhaps the way to summarize it is that the good news is that great
  people like Terance, Ric K. and others are active ANTLR developers
  and ANTLR is evolving.  That's the bad news too when you go to a new
  release, even when it is "new and improved".

  Ian
  www.bearcave.com

 

Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/