🚀 go-pugleaf

In article <1992Apr17.132035.1827@aio.jsc.nasa.gov> bill@epnxt.jsc.nasa.gov writes:
>I am using the UserPath.h and UserPath.m from the Lines program in a  
>program I am writing.  My program works fine if I run it inside my  
>debugger and if I use the NXPing command.  But the same program will not  
>run outside the debugger.  I am at a lost on this one.  Evidently, either  
>something is not being loaded in my program (via #import..but is via the  
>debugger) or something in my program is running away from the window but  
>is forced to slow down in the debugger.

this may not apply to you, but i have problems using user paths under
nextstep 2.0; most people have 2.1 and presumably the problems i am
experiencing are fixed.  in my case, it was not the user paths themselves
which were causing problems, but the very tightly optimized floating point
loops that were filling the user path coordinate array.  when run under
the debugger and single-stepped, they worked perfectly.  when run at full
speed, the arrays were trashed with some spurious results.  the solution
was to edit the assembly language output and insert a nop into the loop to
force the instruction pipeline to empty.

it was all very ugly.