This is the mail archive of the kawa@sources.redhat.com mailing list for the Kawa project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: FAIL format-users test


The stack trace suggests that somehow this.body is null. You could print this.toString() just before the use of body, to get the id field of the LambdaExp, and then add a conditional throw when that LambdaExp is created. It is also possible of of the ExpWalker classes is re-setting body.

Adding --debug-print-expr and --debug-print-final-expr may help track the problem down. If one of the ExpWalker passes is null-ing out the body field, then you could add extra dumps to Compilation.compile.

Of course if it's an implementation bug it may be difficult to track down - if may go away as soon as you add instrumentation!

These days I do most of my Kawa work on a Powerbook G4 running OS X 10.2.6 with Java 1.4.1_01-39, occasionally using GCJ or JDK on Red Hat Linux.

Jim White wrote:

Although I use XSLT a fair bit, I haven't used Kawa's so I don't know what might be affecting it. I don't think it uses Xalan, but if it did then it would be easy to get things mixed up with the various versions and the changes from JDK 1.3 to 1.4.

Kawa's XLST implementation is completely self-contained. It implements XLST just as another Kawa language, compiling an XSLT stylesheet to bytecode. However, it is very incomplete, because I haven't put much time into it. It's more of a proof of concept than anything else.
--
--Per Bothner
per@bothner.com http://per.bothner.com/




Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]