Dear Claus, thank you very kind for your fast answer. I have thought about this solution, but I thought that Hugs would catch that error. But if not, then only Windows is left to do so. Might it be a suggestion towards the designer/programmers of Hugs98? If so to your opinion, would you please pass this on to them. Best wishes, Sander van Tuijl. 9-5-2002 19:25:55, "C.Reinke" <C.Reinke@ukc.ac.uk> wrote:
To the Hugs-Bugs problemsolvers,
The following lines give a fatal error trying to match
with
convect [1] in Winhugs running under windows 98:
convect [] = [] convect (h:t) = convect([h]) ++ convect(t) convect a = a
There is no error message of Hugs itself. The fatal error message comes from windows Hugs is exited by Windows.
haven't seen an answer yet, and it looks fairly obvious, so:
[1] == 1:[], which means that
convect [1] -> convect [1] ++ convect [] -> (convect [1] ++ convect []) ++ convect [] -> ..
the program overflows all bounded resources, typically the stack first. on some platforms, hugs doesn't manage to capture that overflow error, so it (kind of) crashes. which is what you see.
hth, claus