Re: Compiler Construction Tools? (ANTLR)

In article 98-07-183,
    shindle@toocool.com wrote:


> PCCTS:
> tried, true, supported, and apparently fairly popular. Unfortunately,
> our parser guy thinks the C support of the new java based ANTLR needs
> some more testing before we should try it.


One advantage of ANTLR is that the generated source is human readable. I
haven't used the C++ code generator, only the Java one, so I can't comment on
its stability. You can find out more about that on the antlr-interest mailing
list (send "subscribe" to antlr-interest-request@java.magelang.com) Also see
www.antlr.org.


I can say ANTLR is suited exactly to your needs of AST generation,
manipulation and code emitting. I used it to write a six pass translator
from a 4GL (Advanced Revelation) to Visual Basic.


Assuming that SQL embedded in your COBOL is easily recognized by some
beginning and end marker, it would be easy to switch to an SQL parser after
seeing the beginning mark and back to the COBOL after the end mark. They
would be written as separate parsers but the COBOL one would call the SQL
one. There is no global state shared by parsers, so you can have as many
instances of as many different types as you want.


Monty
--

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