Mail Archives: djgpp/1997/02/28/01:25:09
I discovered a peculiar glitch in RHIDE (1.1b).
In an attempt to speed up some code, I manually unrolled a pair of nested
loops that both ran for a constant number of iterations, using macros. The
result is that the macro for the unrolled loop would expand into
ultimately 270 or so (legitimate) lines of C code.
Strangely, the resulting source file (with four such unrolled loops and
very little else) would not compile from in RHIDE but compiled fine from
DOS! It also compiled fine from a DOS shell launched from RHIDE.
The RHIDE error message was especially strange:
Compiling mbkgnd.c
There were errors.
Just like that. No errors, no warnings, just "There were errors" and an
untimely end to the make operation. When I entered the selfsame GCC
command line as RHIDE used, from a DOS shell, it compiled fine. Strangely,
the "End of GCC" message was lacking though -v option was used to try to
spot any signs of trouble. But the .o file was generated correctly.
When I returned to RHIDE with "exit", then ran make, it compiled the other
source files and linked without errors, and it ran perfectly fine too,
including the unrolled loops.
--
.*. Where feelings are concerned, answers are rarely simple [GeneDeWeese]
-() < When I go to the theater, I always go straight to the "bag and mix"
`*' bulk candy section...because variety is the spice of life... [me]
Paul Derbyshire ao950 AT freenet DOT carleton DOT ca, http://chat.carleton.ca/~pderbysh
- Raw text -