Re: [Jmol-developers] Distinction between script error and freezing

[Rolf H. <rh...@fl...>]

Robert Hanson wrote:
> sorry -- pressed SEND prematurely.
> 
> Rolf, please check what I just checked in for Jmol 11.7.14. I think it
> just might work for you.
> 
> 
> -- Out of Memory errors -- "Java heap space"  -- now caught by the
> script processor and image creator.
> 
> -- jmolGetPropertyAsString("errorMessage") returns the error message
> or "null" if there was none.
> 
> Still could be some problems here, but I'm betting this goes a long
> way to solving the problem.
> 
I have tried revision 10376 and it is even worse than before.

When I tried to load a structure into the application that provokes an
"out of memory" error I just got the following error message:

eval ERROR:
----
         load >> "1if0.pdb1.gz" <<

So I could not even recognize that insufficient memory was the problem
here. And although Jmol didn't freeze it also didn't behave well. For
example the automatic screen refresh didn't work any more.

As long as it is not assured that Jmol can work normally afterwards,
freezing might be better (possibly after sending an appropriate error
message).

For my interface it is essential to know what kind of error occured and
which interface command (which is never a single Jmol comand) provoked
it. Only then it is possible to make proper suggestions to the user
about possible solutions.

I currently don't have the time but will later try to explain again why
I would like to have a more defined error tracking system than just
trying to parse the message stream.

Regards,
Rolf