Subject: Re: -current broken?
To: None <rmk@rmkhome.com>
From: Klaus Klein <kleink@reziprozitaet.de>
List: current-users
Date: 03/17/2004 20:14:35
On Wednesday 17 March 2004 19:01, Rick Kelly wrote:
> I've been trying to build -current for several days.
>
> Continual failures with:
>
> cc -O -I/usr4/obj/tools/compat -I/usr4/obj/tools/compat/include 
> -I/usr/src/tools/lint1/../compat -DHAVE_NBTOOL_CONFIG_H=1 
> -D_FILE_OFFSET_BITS=64 -D_NETBSD_SOURCE 
> -I/usr/src/tools/lint1/../../usr.bin/xlint/lint1 -I.
> -DTARGET_OBJFMT_ELF
> -I/usr/src/tools/lint1/../../usr.bin/xlint/lint1/../arch/i386
> -I/usr/src/tools/lint1/../../usr.bin/xlint/lint1/../common   -o lint1
> cgram.lo scan.lo mem1.lo mem.lo err.lo main1.lo decl.lo tree.lo
> func.lo init.lo emit.lo emit1.lo inittyp.lo -L/usr4/obj/tools/compat
> -lnbcompat -lz -lm tree.lo: In function `foldflt':
> tree.lo(.text+0x4496): undefined reference to `__isnand'
> collect2: ld returned 1 exit status

The cause would be the host tool build somehow picking up the
source tree's <math.h>, which it certainly should not.
Please file a PR.


Thanks,
- Klaus