Hi All,

I wanted to send an email to update everyone on the parser work I've
been doing.  I had gotten quite slowed down working to figure out how
all the pointers are set up for the opcodes and instruments and after
a bit of a circuitous route I've arrived back to reuse a number of the
functions in otran.c for setting all of that up.  I do feel now that
I've gotten a much better grasp and would like to at sometime redo
that part of code to be more encapsulated, but for now I feel I've
gotten to know it enough to know it'll do just fine as it is and to
move on.

I feel that now things are getting very interesting.  I have built a
test suite and a testing tool in python (in the NewParser branch this
is all in the tests folder).  The test tool so far is only moderately
useful as parsing and semantic errors are not yet being truly reported
to Csound so return values from running csound often come up 0 when
they should be otherwise.  The test tool though does currently:

-run all tests defined to run
-prints test name and return value
-captures output of all tests and saves in results.txt
-has optional commandline to view the results of a run in a UI using
Tkinter (python's version of Tk, which I chose because Tk is standard
and comes with all versions of python)

I have been adding tests one at a time and introducing a single
feature each time to keep the process of development test-driven and
incremental.

I am looking at parsing and compiling expressions next.  This should
take some time as I have not yet looked through express.c and the old
parser to see what is involved.  Beyond the expressions I am planning
to have functions implement single arguments first to match the old
behavior, but want to make sure that functions will be able to handle
multiple arguments in the future.  (Perhaps too sometime we should
make User-Defined Functions a possibility, or make opcodes with single
return values callable as functions?).

My list of next actions after that are as follow:

-redo bison parser file for instrument 0 to be generic (right now it
has hardcoded values for what kinds of expressions it accepts, when it
needs to match the behavior of the old parser and accept any
statements that have allowable thread)

-handle polymorphic opcodes: I have found the code that does the
polymorphic check and resolution but need to fit it within the new
parser at the right parser stage (should not be too difficult)

-implement gotos and labels

-handle if's, elseif's, and else's

-implement UDO parsing (should be able to reuse instrument parsing
code pretty easily to do this)

-implement macros


I think though that the most difficult parts seem to be past now and
most of the work ahead looks not all too bad.  Doing the test suite
and working by adding one feature for each test has made the work a
lot more focused and easier to move ahead.  There's still a good bit
of work left to do, but I think it's coming along quite well now.

BTW: I am reusing quite a bit of code from the old parser but have
been trying to break things down into smaller functions with
descriptive names and using doxygen style comments.  I am aiming to
make the new parser as easy to understand as I can make it so that
others are more easily able to get up to speed with how Csound works
and get involved with development.

Thanks!
steven

-------------------------------------------------------------------------
Using Tomcat but need to do more? Need to support web services, security?
Get stuff done quickly with pre-integrated technology to make your job easier
Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
_______________________________________________
Csound-devel mailing list
Csound-devel@lists.sourceforge.net