#Manx 5.0a – strike 1
16-Feb-90 11:06:16
Sb: #90198-#Manx 5.0a – strike 1
Fm: Peter Smithem 71001,3663
To: Mike Spille/Manx 71545,1466
I have not found any identifyable bugs yet. I had the compiler go into an
endless loop on me last night while compiling uEmacs 3.10, but when I
restarted the make, the same file compiled with no problems. I had been
changing header files to make the compiler happy and the changes did not
affect thefile which choked the compiler. In fact with all the changes in
the header for other files, this file must have been recompiled five or six
times and only hung the compiler once… strange.
Error # 167 (expression too complex – simplify and recompile) is not
documented anywhere I can find and was emitted by the compiler when it
tried to evaluate the following expresstion:
char name[SIZE]; struct foo {
char n_name[SIZE];
… etc… }; int *iP; struct foo *fP;
name[*iP] = fP->n_name[*iP]; this blows up
While I have in the past seen similar errors (CP/M comiler ~8 years ago?,
expression with several more pointer dereferences), I've never seen it on
such a simple expression. The 3.6 compiler had no problems with exressions
much more complex than this. While declaring an extra temp variable to hold
the index works, I shouldn't have to do this!
<UP ON SOAPBOX>
This is yet another gripe to another software developer about
…Obnoxious compiler version and copyright notices…
Why is it that everyone insists on throwing this information all over the
screen EVERY TIME THE PROGRAM RUNS??? I know what version of what compiler
I'm running. I know who has the copyrights. Why do I have to sift through
all this garbage to see what happened when I ran the program? A simple
command line switch would suffice if I forgot!
<DOWN OFF SOAPBOX>