[Aimmath-developers] Re: Maple 9.5 and problems with WaLLis and AIM

[Manolis M. <m.m...@ed...>]

Altough I can't claim that I solved the problem after lots of reading 
and development I think I am close to a solution.

The solution is a bit specific to WaLLiS but the logic of the Maple 
class in WaLLiS used to be the same as in AIM and I think this is what 
was causing the problem so one could adjust it for AIM as well. I attach 
here the Wallis Maple95 class that has a different purpose than the AIM 
one but perhaps someone can use it.

More work needs to be for AIM, and perhaps lots of things to change. 
Unfortunately I can't test it for AIM at the moment. In addition playing 
with threads and processes on windows is not a nice experience as you 
end up with processes which you can't easily kill. Unfortunately the 
Maple for linux that I am still using is 8.5 (shame!) so I can't test 
it. If noone manages before I find time to do it then I 'll try at some 
point in the near decade or so ...

The main problem is that (I think) there is some significant change on 
the sizes of the messages processed in Maple and this confused the 
Buffer and strains the output before it is really finished. The main 
work around was to have threads that read the output and error buffers.

see more info about that here 
<http://www.javaworld.com/javaworld/jw-12-2000/jw-1229-traps.html> and 
here 
<http://groups.google.com/groups?hl=en&lr=&threadm=WJmdnXDZv4-ncofdRVn-hQ%40adelphia.com&rnum=1&prev=/groups%3Fq%3Druntime%2Bhang%2Bjava%26hl%3Den%26lr%3D%26selm%3DWJmdnXDZv4-ncofdRVn-hQ%2540adelphia.com%26rnum%3D1> 
and here  
<http://www.artima.com/legacy/answers/Nov2000/messages/216.html>and here 
about sockets <http://www.codeguru.com/java/tij/tij0164.shtml> that may 
prove useful

Hope this helps a bit,

Manolis