Prev: char cannot be dereferenced
Next: JAVAMAIL POP3 - Secure Password Authentication
From: Mike Schilling on 16 Jan 2006 13:07

"Luc The Perverse" <sll_noSpamlicious_z_XXX_m(a)cc.usu.edu> wrote in message
news:9sor93xmag.ln2(a)loki.cmears.id.au...
> "Mike Schilling" <mscottschilling(a)hotmail.com> wrote in message
> news:qfGyf.5433$Jd.2992(a)newssvr25.news.prodigy.net...

>>
>> Note that it could be a bug in the JVM even if your application does
>> contain native code; in that case, the first thing to do is remove all
>> the native code and see of the problem goes away. If so, check your
>> native code thoroughly, checking in particular that all JNI calls are
>> correct and that no memory corruption occurs (e.g. that you're not
>> writing past the bounds of an array.) If the problem doesn't go away,
>> it is a JVM error, and you should check for known bugs in that JVM
>> implementation and/or try a different implementation/version.
>
> It is more likely a problem with native code, if there is native code
> there.

I'm recalling a past life in which I used to write a lot of device drivers
for an OS that was not without bugs. Anytime we ran into one, the first
order of business was to reproduce it with no custom drivers loaded. The
situations are quite analogous.


From: Luc The Perverse on 16 Jan 2006 15:42
"Mike Schilling" <mscottschilling(a)hotmail.com> wrote in message
news:shRyf.2491$H71.987(a)newssvr13.news.prodigy.com...
>>> Note that it could be a bug in the JVM even if your application does
>>> contain native code; in that case, the first thing to do is remove all
>>> the native code and see of the problem goes away. If so, check your
>>> native code thoroughly, checking in particular that all JNI calls are
>>> correct and that no memory corruption occurs (e.g. that you're not
>>> writing past the bounds of an array.) If the problem doesn't go away,
>>> it is a JVM error, and you should check for known bugs in that JVM
>>> implementation and/or try a different implementation/version.
>>
>> It is more likely a problem with native code, if there is native code
>> there.
>
> I'm recalling a past life in which I used to write a lot of device drivers
> for an OS that was not without bugs. Anytime we ran into one, the first
> order of business was to reproduce it with no custom drivers loaded. The
> situations are quite analogous.

Perhaps - but keep in mind the nature of the question. This person didn't
even know what the error meant; that doesn't strike me as someone who is
truly challenging the JVM into the realms of the unknown ;)

--
LTP

:)


First  |  Prev  | 
Pages: 1 2
Prev: char cannot be dereferenced
Next: JAVAMAIL POP3 - Secure Password Authentication