Re: cvs commit: src/share/mk sys.mk
On Tue, Feb 20, 2001 at 07:48:31PM +1100, Bruce Evans wrote: > On Tue, 20 Feb 2001, Kris Kennaway wrote: > > > On Tue, Feb 20, 2001 at 05:14:57PM +0900, Jun Kuriyama wrote: > > > I don't know this is local problem on my environment, but "make > > > buildworld" with old make(1) failed if I did not set MACHINE_CPU in > > > /etc/make.conf. So it seems invoked make(1) in > > > src/secure/lib/libcrypto is old one... > > > > Oh, hmm. It seems we do need the bootstrapping code in sys.mk after > > all. > > No. We must fix whatever cause the wrong (old) version of make to be > used. I think it is just that make doesn't actually get bootstrapped > unless the old make is _very_ old (older than the one in 2.2.0). That would be better -- I checked and we don't appear to rebuild make at all as part of the bootstrap process in make world. We can remove the sys.mk hacks once that works. Kris PGP signature
Re: cvs commit: src/share/mk sys.mk
On Tue, 20 Feb 2001, Kris Kennaway wrote: > On Tue, Feb 20, 2001 at 05:14:57PM +0900, Jun Kuriyama wrote: > > I don't know this is local problem on my environment, but "make > > buildworld" with old make(1) failed if I did not set MACHINE_CPU in > > /etc/make.conf. So it seems invoked make(1) in > > src/secure/lib/libcrypto is old one... > > Oh, hmm. It seems we do need the bootstrapping code in sys.mk after > all. No. We must fix whatever cause the wrong (old) version of make to be used. I think it is just that make doesn't actually get bootstrapped unless the old make is _very_ old (older than the one in 2.2.0). Bruce To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: cvs commit: src/share/mk sys.mk
On Tue, Feb 20, 2001 at 05:14:57PM +0900, Jun Kuriyama wrote: > At 20 Feb 2001 07:54:22 GMT, > Kris Kennaway wrote: > > No, MACHINE_CPU is optional. If you don't have it set, you get the > > vanilla C code. So if you don't have it set at all, you'll get C code > > in OpenSSL as it's always been, then the next time you are using the > > updated make(1) and it will set it to i386, which will get you the > > i386 asm code. > > > > Or you can just set MACHINE_CPU immediately and it will build asm code > > on the first pass. > > I don't know this is local problem on my environment, but "make > buildworld" with old make(1) failed if I did not set MACHINE_CPU in > /etc/make.conf. So it seems invoked make(1) in > src/secure/lib/libcrypto is old one... Oh, hmm. It seems we do need the bootstrapping code in sys.mk after all. Kris PGP signature
Re: cvs commit: src/share/mk sys.mk
At 20 Feb 2001 07:54:22 GMT,
Kris Kennaway wrote:
> No, MACHINE_CPU is optional. If you don't have it set, you get the
> vanilla C code. So if you don't have it set at all, you'll get C code
> in OpenSSL as it's always been, then the next time you are using the
> updated make(1) and it will set it to i386, which will get you the
> i386 asm code.
>
> Or you can just set MACHINE_CPU immediately and it will build asm code
> on the first pass.
I don't know this is local problem on my environment, but "make
buildworld" with old make(1) failed if I did not set MACHINE_CPU in
/etc/make.conf. So it seems invoked make(1) in
src/secure/lib/libcrypto is old one...
===> secure/lib/libcrypto
"/usr/src/secure/lib/libcrypto/Makefile", line 62: Malformed conditional
(${MACHINE_CPU:Mi686})
"/usr/src/secure/lib/libcrypto/Makefile", line 62: Missing dependency operator
"/usr/src/secure/lib/libcrypto/Makefile", line 67: if-less else
"/usr/src/secure/lib/libcrypto/Makefile", line 67: Need an operator
"/usr/src/secure/lib/libcrypto/Makefile", line 69: if-less endif
"/usr/src/secure/lib/libcrypto/Makefile", line 69: Need an operator
"/usr/src/secure/lib/libcrypto/Makefile", line 321: Malformed conditional
(${MACHINE_CPU:Mi686} || ${MACHINE_CPU:Mi586})
"/usr/src/secure/lib/libcrypto/Makefile", line 321: Missing dependency operator
"/usr/src/secure/lib/libcrypto/Makefile", line 323: Malformed conditional
(${MACHINE_CPU:Mi386})
"/usr/src/secure/lib/libcrypto/Makefile", line 323: Missing dependency operator
"/usr/src/secure/lib/libcrypto/Makefile", line 326: if-less endif
"/usr/src/secure/lib/libcrypto/Makefile", line 326: Need an operator
make: fatal errors encountered -- cannot continue
--
Jun Kuriyama <[EMAIL PROTECTED]> // IMG SRC, Inc.
<[EMAIL PROTECTED]> // FreeBSD Project
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message
Re: cvs commit: src/share/mk sys.mk
Kris Kennaway <[EMAIL PROTECTED]> wrote:
>
> On Tue, Feb 20, 2001 at 10:02:57AM +0600, [EMAIL PROTECTED] wrote:
>>
>> Kris Kennaway <[EMAIL PROTECTED]> wrote:
>> >
>> > Modified files:
>> >share/mk sys.mk
>> > Log:
>> > Remove bogus setting of MACHINE_CPU here. There is no need for it.
>>
>> But there MUST be at least one setting for
>> MACHINE_CPU for 'make buildworld' to succeed before new 'make'
>> with this variable is in place.
>
> No, MACHINE_CPU is optional. If you don't have it set, you get the
> vanilla C code. So if you don't have it set at all, you'll get C code
> in OpenSSL as it's always been, then the next time you are using the
> updated make(1) and it will set it to i386, which will get you the
> i386 asm code.
>
> Or you can just set MACHINE_CPU immediately and it will build asm code
> on the first pass.
MACHINE_CPU variable MUST be DEFINED for current
'Makefile' in the 'usr/src/secure/libcrypto' be parsed by the 'make'
without internall definition of that variable. (It stops at the first
"${MACHINE_CPU:Mi686}" construct in the line 62).
N.Dudorov
To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message
Re: cvs commit: src/share/mk sys.mk
On Tue, Feb 20, 2001 at 10:02:57AM +0600, [EMAIL PROTECTED] wrote: > > Kris Kennaway <[EMAIL PROTECTED]> wrote: > > > > Modified files: > >share/mk sys.mk > > Log: > > Remove bogus setting of MACHINE_CPU here. There is no need for it. > > But there MUST be at least one setting for > MACHINE_CPU for 'make buildworld' to succeed before new 'make' > with this variable is in place. No, MACHINE_CPU is optional. If you don't have it set, you get the vanilla C code. So if you don't have it set at all, you'll get C code in OpenSSL as it's always been, then the next time you are using the updated make(1) and it will set it to i386, which will get you the i386 asm code. Or you can just set MACHINE_CPU immediately and it will build asm code on the first pass. Kris PGP signature
Re: cvs commit: src/share/mk sys.mk
Kris Kennaway <[EMAIL PROTECTED]> wrote: > > Modified files: >share/mk sys.mk > Log: > Remove bogus setting of MACHINE_CPU here. There is no need for it. But there MUST be at least one setting for MACHINE_CPU for 'make buildworld' to succeed before new 'make' with this variable is in place. I can 'make installworld' after defining 'MACHINE_CPU= i686' in '/etc/make.conf'. Is it deservs some kind of HEADS UP ? N.Dudorov To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
