On Mon, 22 Aug 2005, Mark Taylor wrote:
> On Fri, 19 Aug 2005, Peter W. Draper wrote:
>
> > Hi Mark,
> >
> > thanks for the continuing good work. I've built and committed new
> > sharables libraries for OS X and Windows as these both pass their tests,
> > however, Solaris and AMD64 are crashing around about the STC tests.
> > Solaris says:
> >
> > [junit] Testcase: testReads took 0.927 sec
> > [junit] Caused an ERROR
> > [junit] AST: Error at line 126 in file Channel.c.
> > [junit] astSet: The attribute setting "666612" is invalid for the given CmpFrame. (AST__BADAT)
> > [junit] uk.ac.starlink.ast.AstException: AST: Error at line 126 in file Channel.c.
> > [junit] astSet: The attribute setting "666612" is invalid for the given CmpFrame. (AST__BADAT)
> > [junit] at uk.ac.starlink.ast.Channel.read(Native Method)
> > [junit] at uk.ac.starlink.ast.StcTest.readStc(StcTest.java:78)
> > [junit] at uk.ac.starlink.ast.StcTest.testReads(StcTest.java:44)
> > [junit] at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
> > [junit] at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39)
> > [junit] at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)
> >
> > [junit] Testcase: testChandra took 0.797 sec
> > [junit] Caused an ERROR
> > [junit] AST: Error at line 126 in file Channel.c.
> > [junit] astSet: The attribute setting "666612" is invalid for the given CmpFrame. (AST__BADAT)
> > [junit] uk.ac.starlink.ast.AstException: AST: Error at line 126 in file Channel.c.
> > [junit] astSet: The attribute setting "666612" is invalid for the given CmpFrame. (AST__BADAT)
> > [junit] at uk.ac.starlink.ast.Channel.read(Native Method)
> > [junit] at uk.ac.starlink.ast.StcTest.readStc(StcTest.java:78)
> > [junit] at uk.ac.starlink.ast.StcTest.testChandra(StcTest.java:55)
> > [junit] at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
> > [junit] at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39)
> > [junit] at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)
> >
>
> Hmm, nothing springs to mind. I'll have a closer look if necessary,
> but first, David, does this ring any bells with you? All I'm doing
> here is an astRead( channel ) on Arnold's ChandraSTCResource file.
> Is 666612 anything you recognise?
>
> > and the AMD crash dump (last reported successful test was testResample):
> >
> > C [libjniast.so+0x6068b] Java_uk_ac_starlink_ast_StcCatalogEntryLocation_construct+0x4f
> > j uk.ac.starlink.ast.StcCatalogEntryLocation.construct(Luk/ac/starlink/ast/Region;[Luk/ac/starlink/ast/KeyMap;)V+0
> > j uk.ac.starlink.ast.StcCatalogEntryLocation.<init>(Luk/ac/starlink/ast/Region;[Luk/ac/starlink/ast/AstroCoords;)V+10
> > j uk.ac.starlink.ast.StcTest.testConstructors()V+112
> > v ~StubRoutines::call_stub
> > V [libjvm.so+0x361175]
>
> Doesn't look like the same thing :-(.
>
> > One possible issue is that I'm building against AST 4.1. Any other ideas?
>
> Are you using 4.1 for the platforms that work as well (e.g. OS X) or
> just for the ones that are failing?
I'm using 4.1 for all platforms, so am not totally happy with those that
work. Had another thought, which was perhaps it was also broken under RH9,
but a build-native test for that succeded.
> Also, is the Solaris you're building on 64 bit or 32?
32 bit.
Cheers,
Peter.
|