meuer@umn-cs.cs.umn.edu (Mark V. Meuer) (03/30/91)
I've finally gotten Perl 4.000 to compile and pass tests on our Silicon Graphics Iris running version 3.3 of the operating system. Thanks to all the people who sent me changes they had to make to earlier versions of Perl to get it to work on SGI machines. Thanks should go especially to Patrick Wolfe: His suggestion that I undefine vfork allowed me to break through the major sticking point. Here is the procedure: 1. Run the normal Configure program. Be sure to turn off optimizing ("-O0") and use *perl's* malloc. Other than that, the default answers seem correct. 2. Edit config.sh. Define d_voidsig and undefine d_vfork. 3. There is a strange bug related to the standard include file "sys/file.h" on the Iri. That file defines a macro called "f_next" which is also the name of a field in one of Perl's structures. This is bad. The way around it is to add the following to EVERY file that includes sys/file.h. (Add it right after the #include <sys/file.h> statement): #include <sys/file.h> /* This is a hack to get this to compile on the Iris. You see, the sys/file.h file on the Iri define f_next to be a macro. This causes problems. */ #ifdef f_next #undef f_next #endif /* f_next */ There might very well be a more appropriate place to put this fix, but this will do the job. These changes should make it compile and run the tests successfully. -mark -- Mark Meuer | The Geometry Center | meuer@geom.umn.edu Observing the first baloon ascent in Paris, [Ben] Franklin heard a scoffer ask, "What good is it?" He spoke for a generation of scientists in his retort, "What good is a newly born infant?" -- John F. Kasson
ktl@wag.caltech.edu (Kian-Tat Lim) (03/30/91)
Three minor comments: 1) You only need "#undef f_next" in toke.c. 2) You can use -lmalloc in place of perl's malloc, if you wish. 3) You can "-O2 -Olimit 3000" on everything but perl.c and cmd.c; they can only use "-O1 -Olimit 3000". I'm still trying to track this down. Also, perl -d still seems to die on me. -- Kian-Tat Lim (ktl@wag.caltech.edu, KTL @ CITCHEM.BITNET, GEnie: K.LIM1)