kcwellsch@watmath.waterloo.edu (Ken Wellsch) (09/24/88)
Trying to compile GCC 1.28 on a Sequent Symmetry (i386 based) (same problem with 1.26 and 1.27): config.h -> config-i386.h tm.h -> tm-seq386.h md -> i386.md aux-output.c -> output-i386.c Offending event extracted from tail of Make output: | ... | stage2/gcc -g -O -Bstage2/ -c tree.c | tm-bsd386.h:202: warning: ASM_OUTPUT_DOUBLE redefined | Assembler: | "tree.c", line 1796: "den0x00000000000000001000" is not followed by a ':' | for a label definition What this bizarre double constant (should be ".double 0d1.00000e+00") means (from the local printf man page): | PRINTF(3S) DYNIX Programmer's Manual PRINTF(3S) | | NAME | printf, fprintf, sprintf - formatted output conversion | | SYNOPSIS | #include <stdio.h> | | printf(format [, arg ] ... ) | char *format; | | ... | | If the exponent is zero and the mantissa is not equal to | zero, then the output is | | [-]den0xdddddddd | | where 0xdddddddd is the hexadecimal representation of the | leftmost 32 bits of the mantissa. | | ...
rms@WHEATIES.AI.MIT.EDU (Richard Stallman) (09/26/88)
I don't expect to do anything about this problem except wait for someone to send me a convincing fix. Chances are only a person with one of those machines can find one. That info about printf explains the last stages of the problem, but someone still needs use a debugger to get to the root of the matter.