🚀 go-pugleaf

In article <gnusenet-1krf5mEigf@uni-erlangen.de>, eilts@late.e-technik.uni-erlangen.de (Hinrich Eilts) writes:
|> I found a bug in compiling epoch-4.2/src/filelock.c with gcc-2.3.3 on a
|> NeXT:
|> In lisp.h is a redefinition of the macro NULL (what a bad idea!), but in the
|> later (over a include-chan started by #include <errno.h>) included
|> /usr/local/lib/gcc-lib/next/2.3.3/include/stddef.h this will be overruled
|> by a new define of NULL (to it's usual value).
|> One of the affected sourcecodelines (line 118) in filelock.c looks like:
|>     if (!NULL (subject_buf)
|> before preprocessing, and after the wrong turn:
|>     if (!((void *)0)  (subject_buf)
|> and it should be:
|>     if (! ((subject_buf)  == (Qnil) )
|> My propositions are: Change of the redifined NULL in epoch to EPOCH_NULL or
|> whatever and a warning-message from gnu-cpp for any redefinitions.
|> 
|> -- 
|>               Bye                                 | G i b   D O S |
|>                    Hinrich Eilts                  | k  e  i  n  e |
|>   (e-mail: eilts@late.e-technik.uni-erlangen.de)  | C h a n c e ! |
|> 

I got around this annoying problem by using sed to replace the NULL
with NIL, which is #define'd to be the same as NULL.  I agree though,
the use of NULL was not a well-thought-out decision.

-- 
+---------------------------------------------------------------+
| Mike D. Kail                        mdk@pegasus.udev.cdc.com  |
| Control Data Systems, Inc.                                    | 
| 4201 North Lexington Avenue            (612) 482-2942 (W)     |
| Arden Hills, MN  55126                 (612) 628-9132 (H)     |
+---------------------------------------------------------------+
| "Nothing great was ever achieved without enthusiasm."         |
|  - Ralph Waldo Emerson                                        |
+---------------------------------------------------------------+