1) My version is now passing xsltc.conf. I haven't run all the other
XSLTC tests for complete regression testing, but this resolves the known
issue of moving lookahead token identification into the RE patterns
rather than hand-coded manipulation of (no longer exposed) buffer
contents. I haven't attempted to compare efficiency with the
hand-coded-lookahead disambiguation, but frankly I don't think it
matters much; XSLTC is more about execution speed than compilation
speed, and compiling it into the core tokenizer may actually be more
efficient.
2) Samples is now a sub-project. They compile, though I need to clean up
the imports so we don't get double-inclusion warnings (though those are
usually harmless). Now that XSLTC should be working, I may be able to
test those, though they aren't well documented.
I have not yet replicated:
* Javadoc/page generation.
* "Production/distribution" aggregated zipfiles and jarfiles.
I still need to look at
* Moving the cup/lex files from src/main/java/ to src/main/grammar/ or
resources/ or something like that.
* What else should be moved into resources/?
* Are there other sub-projects needed?
* Mavenizing xalan-test and moving it into a sub-project of xalan-java
* Running a _complete_ regression suite
* Getting someone more experienced with Maven to detect and correct
violations of Maven idiom/conventions/cleanliness.
After those we can discuss merging any late changes from the main branch
and cutting over officially. Note that since this reorganizes
directories the merge may be a bit more manual than ideal; Oh Well. (And
apologies to Mukul for the hassle of moving things between directories.)
"Please stand by. We are making mistakes as efficiently as possible."
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@xalan.apache.org
For additional commands, e-mail: dev-h...@xalan.apache.org
