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

Re: Error recovery


Error recovery:

> > please do fetch the latest CVS source tree, build the sablecc.jar and try
> > it out with your stuff which does use the error recovery features.
> So, is it working for you?  If any of you has a comment, please send it
> on the mailing-list.

Yup, it works like a dream ;o)
I have just commited some older changes, which I suspect did not make into
CVS by mistake yet, together with some more recent changes. Below is the
CVS log info. It should all compile and work with all the example

Would be good if Dan could test it with his error recovery "client" code.
Later we should/could add error recovery to some of the example grammars.

Mailing lists:
Regarding the use of mailing lists, maybe we should all move and
consistently use the SourceForge mailing lists? We could go into more
detailed discussions on developers mailing list, and have more general
discussions on users one?

Java interpreter:
I am working on a Java/scripting interpreter - I hope to make it full Java
interpreter at some stage, but at the moment I just need a Java expression
evaluation tool.
It does use SableCC (of course) and makes use of this famous "hashtable
pattern" for traversing Java AST (of course)  ;o)
It is different from bsh or other java scripting frameworks: it builds an
explicit complete internal model of the expressions, blocks, and in the
future classes, methods, etc. It means it will be slower and more memory
consuming, but will enable reification of expressions, methods, classes,
references, from within the Java code itself. One will be able not only
inspect a class, but change the behaviour, add methods, change method
implementations, etc.  It does simple things at the moment, and I may put
it on sablecc as a subproject or as a seperate SF project if more people
would be interested in using/helping/developing it. Please, let me know.

best regards

CVS Log Message:
*) Fixing the Executable HashSet bug (the order of the command line
arguments was dependent on the HashSet implementation; changed to use List

*) Fixing the Error Recovery code generation to be dependent on the Parser
generation (no production -> no Parser generation -> no error recovery).
Commiting some older error recovery fixes.

*) Setting up build.xml and SableCC.java to explicitely report exceptions
(no optimizations for line numbers and printStackTrace() calls). Should be
changed back to optimized and "quiet" code for the release.
All compiles and works with the grammars examples.