[Proj] Problem with program on Windows

support.mn at elisanet.fi support.mn at elisanet.fi
Thu Aug 11 17:40:41 EST 2011


I agree that it would be an additional labor but the Proj.4
grammar is so easy that it does not take very many lines
to implement and of course you could always exclude the
scanner from your project if that was not needed since
that would just be some kind of a preprocessor. It would
greatly improve the quality of the definition data and help
users to avoid most obvious mistakes. It would also make
the processing otherwise more simple since the code
would not need to do any repeated checking.

I did also several compiler and language projects in 80's
and 90's even one full relational database language.



Frank Warmerdam [warmerdam at pobox.com] kirjoitti: 
> On Thu, Aug 11, 2011 at 2:53 PM,  <support.mn at elisanet.fi> wrote:
> > Frank,
> >
> > that is the reason I have been talking about the syntax check
> > to be connected to Proj.4 ... that would then give at least a warning
> > (if switched on) that the user is giving meaningless input to Proj.4
> > and it would handle situations like this here (and maybe several
> > similar later) and minimize the input data garbage.
> Janne,
> It would be desirable, but it isn't something I intend to work on in
> the near term.  I would note doing it properly would require
> someone *knowing* the grammar.  In fact I am often not aware
> of the syntax of many of the esoteric options and I'd hate to
> break stuff that is working for some people.
> > Syntax scanning is very easy to implement using "Recursive
> > descent parser" and will not yield any massive source code.
> >
> > http://en.wikipedia.org/wiki/Recursive_descent_parser
> I have written lots of parsers, sometimes using parser generators
> and sometimes by hand.  Depending on the interpretation of
> "massive" I do not entirely agree with your statement.  I believe
> having a proper parser as part of PROJ.4 would non-trivially
> bulk up the project.

More information about the Proj mailing list