[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: cppx



Hi Steve,

I was playing around with your test program "vorpalsword.cpp" and have gotten cppx to work on it. I did some major surgery so the program itself is broken.

What I did was take out the iostream library. This meant that I had to fake or cut out sections in your code.

I did this is /home/sesweb/yuzok/steve and you can see vorpalsword.cpp there. Look at all the //Yuzo: comment areas, that's where I changed things.

Note that err2 is the cppx run that created the gxl file, there are a couple of compiler warnings in it, but cppx continued after then. Note the size of the gxl! 38566 nodes!

If this is typical for the doc sizes will be dealing with then we may have to revisit the IMR and data bindings as it may really take a bit out of the Java interpreter.

Yuzo

--On Wednesday, July 28, 2004 12:25 AM -0700 Stephen Regan <sregan@uci.edu> wrote:


Hi Yuzo,

I installed the fact extractor cppx in the tools directory in sesweb.
I've tried running a c++ file through it, but I'm getting a lot of parse
errors.  The file I tried does compile using g++ (it's an old homework
assignment I did in Visual Studio).  I'm going to try using some of the
example files on the cppx web site.  I believe I have cppx set up
correctly.  There is a script file in the bin directory that I modified
(going by the installation instructions).  I think the problem I'm having
has more to do with the preprocessing done by cppx, which uses GNU gcc.
The script file is called 'cppx' and is in
'/home/sesweb/tools/cppx/cppx/bin.'  To run cppx just type 'cppx
filename'.  The default setting is to output a file 'filename.gxl'.  To
get a ta file requires (from what I gather at this point) that the 'cppx'
script be modified.  In case you're interested here is a link with some
info about cppx: http://swag.uwaterloo.ca/~cppx/doc/cppx/arch.html and
here are some simple c++ samples with corresponding ta output (the gxl
links are dead :(   )
http://swag.uwaterloo.ca/~dean/cppx/examples/expr/index.html.  I'm going
to fiddle with this some more, I'm just sending this message because you
were working with a fact extractor.  I'm also getting yet another
extractor up and running, Columbus/Can, that works on Windows.  More on
that story as it develops.

Steve