Re: coding conventions (Steven Bosscher)
24 Feb 2003 13:51:15 -0500

          From comp.compilers

Related articles
coding conventions (2003-02-05)
Re: coding conventions (2003-02-11)
Re: coding conventions (2003-02-21)
Re: coding conventions (2003-02-24)
| List of all articles for this month |

From: (Steven Bosscher)
Newsgroups: comp.compilers
Date: 24 Feb 2003 13:51:15 -0500
References: 03-02-009 03-02-065
Keywords: yacc, practice
Posted-Date: 24 Feb 2003 13:51:15 EST (Alex Colvin) wrote in message news:03-02-065...
> >I'm fairly new to this flex/bison stuff, and I was wondering - is
> >there a recommended coding convention for flex/bison ? like for
> >example should I have named the "subpart" differently, etc.
---- 8< ----
> I don't have any formal style guide, but I'll push the convention that the
> yacc actions should be kept as short as possible, either an assignment or
> a function call. Not only does long precedural code make the grammar hard
> to follow, but you don't want to put anything you might need to debug in
> your *.y file.

I agree that the yacc actions should be short, because it keeps the
grammar clean and readable. I don't see the debugging problem though.
  Can you explain that a but further please?

Last time I has to work with a grammar that was stuffed with
production actions, I could step through the code in the .y file in
the debugger. In fact in was quite helpful (but still ugly) that the
code was in the grammar because you can see which production you were


Post a followup to this message

Return to the comp.compilers page.
Search the comp.compilers archives again.