This is the mail archive of the kawa@sourceware.org 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: Patch: new special form #!native


On Sep 29, 2011, at 2:32 PM, Per Bothner wrote:

This looks clean, but it causes a regression (tested using jdk7):

testsuite$ make check-knucleotide-2
CLASSPATH="/home/bothner/JWebPane/jfx/artifacts/sdk/rt/lib/ jfxrt.jar:/home/bothner/Java/servlet.jar:.:$CLASSPATH" \
../bin/kawa.sh --warn-undefined-variable --warn-invoke-unknown- method --warn-as-error --main -T knucleotide2 -C ./knucleotide-2.scm
(compiling ./knucleotide-2.scm to knucleotide2)
./knucleotide-2.scm:13:1: missing implementation for java.util.Map $Entry.hashCode()int
make: *** [knucleotide2.class] Error 255

Bummer. Confirmed with 1.6.


I assume this is because of the changes to ClassExp#compileMembers - probably the second change.
Somehow the (native) implement of java.lang.Object#hashCode is skipped. Probably easy to fix.

That must be it. Reverting the second change (but nothing else) fixes it,
and my #!native test (which requires jna or jnaerator) still works.


To be honest, I'm really not sure what that section of ClassExp is doing;
I just did a search for isAbstract() and added isNative() calls in the
vicinity assuming that in most cases that would be a reasonable thing to
do.


Does it otherwise seem like a reasonable addition?

-Jamie

--
Jamison Hope
The PTR Group
www.theptrgroup.com




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