abashere

As far as 'db's are concerned, that is true. I don't remember the port that wanted 'db'. Using the shotgun approach, I probably loaded all the packaages. Then looked at the ./configure test and fould that configuration allowed the ./configure to work. It was the result of a direct copy of that include file (db.h from db3(?)) into /usr/include.


I think I have to diagree based upon the recorded scripts... see below


Not the problem; simply doing the copy as mentioned above solved the compile problem.


In the log below, there is the line:
cc -O -pipe -nostdinc -I/usr/include -I. -I/usr/src/sys/dev/aic7***/aicasm

Which says to me:
-O optiize
-pipe pipe between stages
-nostdinc no standard default include path
-I/usr/include search /usr/include first for files !!!!!
-I. search "current" file for includes
-I/usr/src/sys/dev/aic7***/aicasm search source directory.

The source has:

#ifdef __linux__
#include "aicdb.h" #else
#include <db.h>
#endif ...
Which <db.h> do you think got loaded?

It looks like "/usr/include/db.h" to me. One of the files named 'db.h' has the missing R_FIRST (if not replaced by a different db.h from a different package or port).

Andrew Bashere
-------- from /var/tmp/mw.0703
===> vesa
rm -f setdef0.c setdef1.c setdefs.h setdef0.o setdef1.o vesa.ko vesa.kld vesa.o scvesactl.o @ machine symb.tmp tm
p.o opt_vga.h
rm -f .depend GPATH GRTAGS GSYMS GTAGS
cd /usr/obj/usr/src/sys/HOTSPARE; MAKESRCPATH=/usr/src/sys/dev/aic7***/aicasm make -DBOOTSTRAPPING -f /usr/src/
sys/dev/aic7***/aicasm/Makefile
Warning: Object directory not changed from original /usr/obj/usr/src/sys/HOTSPARE
cc -O -pipe -nostdinc -I/usr/include -I. -I/usr/src/sys/dev/aic7***/aicasm -D__FBSDID=__RCSID -c /usr/src/sys
/dev/aic7***/aicasm/aicasm.c
cc -O -pipe -nostdinc -I/usr/include -I. -I/usr/src/sys/dev/aic7***/aicasm -D__FBSDID=__RCSID -c /usr/src/sys
/dev/aic7***/aicasm/aicasm_symbol.c
/usr/src/sys/dev/aic7***/aicasm/aicasm_symbol.c: In function `symbol_delete':
/usr/src/sys/dev/aic7***/aicasm/aicasm_symbol.c:91: warning: passing arg 2 of pointer to function from incompatib
le pointer type
/usr/src/sys/dev/aic7***/aicasm/aicasm_symbol.c:91: too few arguments to function
/usr/src/sys/dev/aic7***/aicasm/aicasm_symbol.c: In function `symtable_open':
/usr/src/sys/dev/aic7***/aicasm/aicasm_symbol.c:135: warning: assignment makes pointer from integer without a cast
/usr/src/sys/dev/aic7***/aicasm/aicasm_symbol.c: In function `symtable_close':
/usr/src/sys/dev/aic7***/aicasm/aicasm_symbol.c:151: structure has no member named `seq'
/usr/src/sys/dev/aic7***/aicasm/aicasm_symbol.c:151: `R_FIRST' undeclared (first use in this function)
/usr/src/sys/dev/aic7***/aicasm/aicasm_symbol.c:151: (Each undeclared identifier is reported only once
/usr/src/sys/dev/aic7***/aicasm/aicasm_symbol.c:151: for each function it appears in.)

.......
/usr/src/sys/dev/aic7***/aicasm/aicasm_symbol.c:486: `R_FIRST' undeclared (first use in this function)
/usr/src/sys/dev/aic7***/aicasm/aicasm_symbol.c:487: structure has no member named `seq'
/usr/src/sys/dev/aic7***/aicasm/aicasm_symbol.c:524: `R_NEXT' undeclared (first use in this function)
*** Error code 1

Stop in /usr/obj/usr/src/sys/HOTSPARE.
*** Error code 1

Stop in /usr/src.
*** Error code 1

Stop in /usr/src.
</usr/src>#
---------------------------------

jens schweikhardt

#
# # Jens Schweikhardt wrote:
#> Andrew Bashere <abashere@yahoo.com> wrote
#> in <fd707556.0307080529.3c7b056a@posting.google.com>:
#> # keywords: failed kernel aic7***x aicasm_symbol.c R_FIRST R_LAST db.h
#> #
# # ...snip... #
#> # that wanted some "db". Unfortunately, there is a db2, db3, db4, and
#> # ndb and gdb.
#>
#> Something on your machine appears seriously misinstalled.
#
# As far as 'db's are concerned, that is true. I don't remember the port that wanted 'db'. Using the shotgun approach, I probably loaded all the packaages. Then looked at the ./configure test and fould that configuration allowed the ./configure to work. It was the result of a direct copy of that include file (db.h from db3(?)) into /usr/include.
#>
#> # It is odd that the kernel compile depends upon the existing (old)
#> # system files to be correct to compile correctly.
#>
#> In fact the kernel does not depend on any source outside /usr/src/sys.
#
# I think I have to diagree based upon the recorded scripts... see below
#>
#> # One would think that
#> # it should be looking at what will be installed leter during that
#> # install process.
#> #
#> # I simply copied /usr/src/include/db.h into /usr/include/db.h and the
#> # build completed.
#>
#> Then something else changed, as a kernel compile does not pull in
#> anything from /usr/include.
# See below ....
# cc -O -pipe -nostdinc -I/usr/include -I. -I/usr/src/sys/dev/aic7***/aicasm

You are right, I just looked at the logs of the last current run.
-I/usr/include does appear only in the aic7*** driver. Hmmm.

# #ifdef __linux__
# #include "aicdb.h" # #else
# #include <db.h>
# #endif # ... #
# Which <db.h> do you think got loaded?
#
# It looks like "/usr/include/db.h" to me. One of the files named 'db.h' has the missing R_FIRST (if not replaced by a different db.h from a different package or port).
#
# Andrew Bashere

AFAIK, a buildkernel is supposed to be selfcontained under /usr/src. The
correct db.h to use would be /usr/src/include/db.h so it appears to me
the build process has a bug where -I/usr/include should be
-I/usr/src/include instead. I'll ask the aic maintainer.

Regards,

Jens
--
Jens Schweikhardt http://www.schweikhardt.net/
SIGSIG -- signature too long (core dumped)

g. remlin

Saw this, too.
What I did was go up to /usr/src and:

% make kernel

which built it. But, yes, it used to
be true that you didn't need to build
a kernel there.