On 27 Jul 97 at 12:00, Peter Gerwinski wrote:
There is one remaining bug I want to fix before I release the next beta, the GPI problem reported by Bill Currie. (See our recent mail on this list.) Unfortunately, I cannot reproduce this bug on any platform I have access to. If somebody else can reproduce it and help to find out under what circum- stances the error occurs and when it doesn't, it would be great help.
Yesterday, I tried out the diff you gave me, but it didn't fix the gpi problem :( I've noticed it only ever happens when writing out the gpi file after encountering the `implementation' keyword. As both bzeroing the array and moving the array to the heap don't fix the problem, I strongly suspect memory is being freed without giving its month's notice ;)
Couldn't get my reply to you earlier due to local network problems.
Bill -- Leave others their otherness.