Skip to contentSkip to conversation

Uncompilable....

DISCUSSION Harry Kalogirou 4 participants 5 messages Jun 6, 2002

lore
Share & tools
lore
#0 lore raw

make -C . all
make[1]: Entering directory `/home/harkal/cvs/elks'
make -C kernel kernel.a
make[2]: Entering directory `/home/harkal/cvs/elks/kernel' bcc -0 -i -O -I/home/harkal/cvs/elks/include -DELKS_VERSION_CODE=0x00010002 -DUTS_RELEASE=\"0.1.1-pre2\" -D__KERNEL__ -ansi -c -o version.o version.c
version.c:10.48: error: string longer than dimension make[2]: *** [version.o] Error 1
make[2]: Leaving directory `/home/harkal/cvs/elks/kernel' make[1]: *** [kernel/kernel.a] Error 2
make[1]: Leaving directory `/home/harkal/cvs/elks'
make: *** [elks] Error 2

That is the output of making the current CVS tree.

Harry

#1 lore raw
Quoted 9 lines · Harry Kalogirou

bcc -0 -i -O -I/home/harkal/cvs/elks/include

(quoted diff) +0 -2
-DELKS_VERSION_CODE=0x00010002 -DUTS_RELEASE=\"0.1.1-pre2\" -D__KERNEL__
-ansi -c -o version.o version.c

version.c:10.48: error: string longer than dimension make[2]: *** [version.o] Error 1
make[2]: Leaving directory `/home/harkal/cvs/elks/kernel' make[1]: *** [kernel/kernel.a] Error 2
make[1]: Leaving directory `/home/harkal/cvs/elks'
make: *** [elks] Error 2

I compiled it about 5 minutes ago -and it built fine ?

#2 lore raw
Quoted 12 lines · Alan Cox
Quoted 9 lines · Harry Kalogirou

bcc -0 -i -O -I/home/harkal/cvs/elks/include

(quoted diff) +0 -2
-DELKS_VERSION_CODE=0x00010002 -DUTS_RELEASE=\"0.1.1-pre2\" -D__KERNEL__
-ansi -c -o version.o version.c

version.c:10.48: error: string longer than dimension make[2]: *** [version.o] Error 1
make[2]: Leaving directory `/home/harkal/cvs/elks/kernel' make[1]: *** [kernel/kernel.a] Error 2
make[1]: Leaving directory `/home/harkal/cvs/elks'
make: *** [elks] Error 2

I compiled it about 5 minutes ago -and it built fine ?

Hmm broken in a different way here, I just slurped the cvs and did a default .config

(cd ../.. ; ld86 -0 -i   \
	arch/i86/boot/crt0.o arch/i86/boot/crt1.o \
	init/main.o kernel/kernel.a fs/fs.a lib/lib.a net/net.a
fs/minix/minixfs.a arch/i86/kernel/akernel.a arch/i86/lib/lib86.a
arch/i86/mm/mm.a  arch/i86/drivers/char/chr_drv.a
arch/i86/drivers/block/blk_drv.a \
	-t -M -o arch/i86/boot/system > System.tmp ; \
	sort -k4 System.tmp > System.map ; rm -f System.tmp ) 

undefined symbol: _rs_ops

This seems to be defined in arch/i86/drivers/char/serial.c

but inside a:

#ifdef CONFIG_CONSOLE_SERIAL

but it's used in ntty.c within a

#ifdef CONFIG_CHAR_DEV_RS

Paul

#4 lore raw

Hi Harry.

Quoted 5 lines
Quoted 1 line

I compiled it about 5 minutes ago -and it built fine ?

Well it has to do with the UTS_VERTION define that is probably generated at the compliling machine. This problem was encountered again some time ago....

I remember well the first time this problem came up, and I committed a patch to fix it then. Somewhere in between, that patch was obviously reverted, so I've just re-committed it along with a patch to display the contents of include/linuxmt/compiler-generated.h (the header file that is created as part of the compile procedure) to make such obvious.

I've checked that the result both compiles and boots, so no problems there either - now it's bed for me so I can be ready for the morning...

Best wishes from Riley.