[armedbear-devel] debug stack overflow? (to track down regression)

[Alan Ruttenberg]

On Tue, Apr 27, 2010 at 2:40 PM, Erik Huelsmann <ehuels at gmail.com> wrote:
> Hi Alan,
>
> On Tue, Apr 27, 2010 at 4:31 AM, Alan Ruttenberg
> <alanruttenberg at gmail.com> wrote:
>> Hi,
>>
>> This (a call in my LSW system) used to work for me and is now failing
>> in trunk. However I don't seem to have anything to go on in the stack
>> trace. Any suggestions?
>
> Right. What you're seeing is expected: the stack gets unwound before
> raising this error, because there's no room left on the stack to
> actually handle it. It's a bit unfortunate.

Any chance of writing the stack to a file before tossing it? Is there
a way of signaling the stack condition when there's still some stack
left - is that a knob that Java offers?

>
> What I do to work on stuff like this is to make sure I get visual
> feedback (printf) to see where in the file the error might occur. I'm
> not sure if I can come up with anything else.

So I've tried to do that, but it looks like it's in or around a call
to a java function. I can try to dissect it more by printing inside my
jss code.

Bleh.

-Alan


>
> Bye,
>
>
> Erik.
>