compile problem under -current on i386 #1

[Marc Espie]

On Wed, Jan 26, 2000 at 10:16:53PM +0100, Matthieu Herrb wrote:
> Hi,
>
> running 'make build' on -current i386 yield to the following errors
> with my source tree, in sys/arch/i386/stand/libsa:
>
> /usr/src/sys/arch/i386/stand/libsa/gidt.S: Assembler messages:
> /usr/src/sys/arch/i386/stand/libsa/gidt.S:214: Error: Rest of line
> ignored. First ignored character is `d'.
> /usr/src/sys/arch/i386/stand/libsa/gidt.S:214: Error: Rest of line
> ignored. First ignored character is `d'.
What does your /usr/bin/cpp look like ?
It should be a shell-script, the real cpp gcc does install lives in
/usr/libexec, and the wrapper does -traditional, specifically to avoid
this kind of problem...

It's likely you did a make install by hand somewhere, forgetting about
a Makefile.bsd-wrapper...
--
Marc Espie
|anime, sf, juggling, unicycle, acrobatics, comics...
|AmigaOS, OpenBSD, C++, perl, Icon, PostScript...
| `real programmers don't die, they just get out of beta'

[Matthieu Herrb]

Hi,

running 'make build' on -current i386 yield to the following errors
with my source tree, in sys/arch/i386/stand/libsa:

/usr/src/sys/arch/i386/stand/libsa/gidt.S: Assembler messages:
/usr/src/sys/arch/i386/stand/libsa/gidt.S:214: Error: Rest of line
ignored. First ignored character is `d'.
/usr/src/sys/arch/i386/stand/libsa/gidt.S:214: Error: Rest of line
ignored. First ignored character is `d'.

[and so on]

The following patch takes care of it, so I assume cpp has changed its
behaviour in 2.95.2 ?

Index: gidt.S
===================================================================
RCS file: /cvs/src/sys/arch/i386/stand/libsa/gidt.S,v
retrieving revision 1.23
diff -u -r1.23 gidt.S
--- gidt.S 1998/09/25 06:48:52 1.23
+++ gidt.S 2000/01/26 21:06:42
@@ -204,8 +204,8 @@
lidt Idtr
ret

-#define IPROC(n) X/**/n
-#define IEMU(n) IPROC(emu/**/n)
+#define IPROC(n) X##n
+#define IEMU(n) IPROC(emu##n)
.align 3
idt:
#define idte(e) \
@@ -221,7 +221,7 @@
/* Maskable interrupts (32-255) */
/* BIOS entry points (32-63) */
/* DOS entry points (64-80) */
-#define idtb(b) idte(emu/**/b)
+#define idtb(b) idte(emu##b)
idtb(0); idtb(1); idtb(2); idtb(3); idtb(4); idtb(5)
idtb(6); idtb(7); idtb(8); idtb(9); idtb(10); idtb(11)
idtb(12); idtb(13); idtb(14); idtb(15); idtb(16); idtb(17)

--
Matthieu

[Matthieu Herrb]

You wrote (in your message from Wednesday 26)

> What does your /usr/bin/cpp look like ?
> It should be a shell-script, the real cpp gcc does install lives in
> /usr/libexec, and the wrapper does -traditional, specifically to avoid
> this kind of problem...
>

You're right, that's the problem. Thanks.

> It's likely you did a make install by hand somewhere, forgetting about
> a Makefile.bsd-wrapper...

Probably, but I only remember to have run the following commands in
gnu/egcs/gcc to build and install egcs 2.95.2 before starting make
build, and it left me with the real cpp in /usr/bin:

make -f Makefile.bsd-wrapper obj

make -f Makefile.bsd-wrapper cleandir \
make -f Makefile.bsd-wrapper depend | twice
make -f Makefile.bsd-wrapper |
make -f Makefile.bsd-wrapper install /

--
Matthieu