[antlr-interest] Re: caching AST and parsed information

[daniel_raichle]

Ok, that's right it won't help if I'm parsing the .c files which still
contain macros. I didn't think abaout that.

But the other approach sounds very interesting. How would I have to
change the lexer so that reads through the whole section and builds
the checksum before passing the tokens to the parser? What about
typedefs in this section, would I have to cache the symbol table, too?

Thank you for your help!

Daniel

--- In antlr-interest at yahoogroups.com, mzukowski at y... wrote:
> This is certainly possible.  Of course this won't help much if your
.c files
> actually use the macros defined in the temp.h file.  But it's easy
to write
> a rule that parses the #include and then builds a whole copy of the
> previously parsed tree.  
> 
> Another approach would be to recognize the #file directives and do
something
> smart like parse and checksum the whole section contributed by 
"temp.h"
> When lexing the next file first just read thru and checksum the
section and
> if it already exists just use it out of the cache (from the
parser--just
> pass a special token back from the lexer which represents the whole
section
> and which the parser will replace with the cached tree.)
> 
> Monty
> 
> -----Original Message-----
> From: daniel_raichle [mailto:daniel.raichle at e...] 
> Sent: Tuesday, January 20, 2004 1:38 AM
> To: antlr-interest at yahoogroups.com
> Subject: [antlr-interest] caching AST and parsed information
> 
> Hi,
> 
> I am using antlr to instrument multiple (small) C-files
automatically
> with Monty's C-grammar.
> 
> Now I have the problem that all of my C-files include a file called
> "temp.h" in which some macros are defined and which includes some
> other project specific .h-files and some standard .h-files (eg.
> stdio.h, math.h, ...). Because of the macros and typedefs in the
> project-specific .h-files I am using a preprocessor before parsing
> with antlr and I am getting very large files because of all the
> includes, mainly the standard .h-files. Therefore parsing those
files
> is very slow.
> 
> Now my thoughts are: as all of my C-files only include this single
> .h-file, I could possibly parse the temp.h file first and then cache
> all the informations, antlr has parsed up to this point. After that
I
> could parse my c-files directly without using a preprocessor before
> and ignoring the #include "temp.h"-directive.
> 
> How could I do that? Is this possible with antlr?
> 
> Thanks,
> Daniel Raichle
> 
> 
>  
> 
> Yahoo! Groups Links
> 
> To visit your group on the web, go to:
>  http://groups.yahoo.com/group/antlr-interest/
> 
> To unsubscribe from this group, send an email to:
>  antlr-interest-unsubscribe at yahoogroups.com
> 
> Your use of Yahoo! Groups is subject to:
>  http://docs.yahoo.com/info/terms/


 

Yahoo! Groups Links

To visit your group on the web, go to:
 http://groups.yahoo.com/group/antlr-interest/

To unsubscribe from this group, send an email to:
 antlr-interest-unsubscribe at yahoogroups.com

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