Re: generated code

Thank you for your long response. I am satisfied, tha you confirm my
thesis about the advantages of the code created by recursive descent
LL compiler compilers:


>For most people, your thesis holds true


Then you try to show the restrictions of this method. Though I don't
really understand what's meant by


>we run Yacc++ engines in a "call-back" mode designed for being used
>in an event driven loop.


I absolutely agree, that direct processing of the input often is not
sufficient.


I try to take the perspective of a user of a parsing tool, who mostly
isn't an experienced compiler theorist. E.g. he might want to take
some contents from some Internet pages and write programming code,
which feeds them into an stl container. For such tasks not only a
parser can be constructed with the TextTransformer in minutes, but you
even get the result. This, due to the simple insertions of -
interpreted - semantic code into the parsing productions. The
TextTransformer interpreter uses a TextTransformer parser and can


> handle loops (or conditional statements)


Indeed the interpreter uses an AST. Though I'm concentrated on
practical application of compiler tools, I am also quite ambitious to
extend the capabilities of the TextTransformer as far as
possible. Until now I found no real task that I could not in principle
fulfill with this tool. But I confess, until now I stayed


>in the cocoon of the single threaded top-down world.


Thanks for your attempt to show me the world outside.
--
mailto:Meyer-Eltz@t-online.de


url: http://www.texttransformer.de
url: http://www.texttransformer.com

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