Re: [openssl.org #410] Re: HP-UX build problems with 0.9.6h

2002-12-31 Thread Marko Asplund via RT 

On Mon, 23 Dec 2002, Andy Polyakov wrote:

> > i tried this but linking the openssl program as well as the tests fails
> > with the following message:
> >   cc -o openssl -DMONOLITH -I.. -I../include  -DOPENSSL_THREADS 
>-D_REENTRANT -DDSO_DL -DOPENSSL_NO_KRB5 -DOPENSSL_NO_ASM +DA2.0 +DS2.0 +O3 
>+Optrs_strongly_typed +Olibcalls -Ae +ESlit -DB_ENDIAN -DMD32_XARRAY openssl.o 
>verify.o asn1pars.o req.o dgst.o dh.o dhparam.o enc.o passwd.o gendh.o errstr.o  ca.o 
>pkcs7.o crl2p7.o crl.o  rsa.o rsautl.o dsa.o dsaparam.o  x509.o genrsa.o gendsa.o 
>s_server.o s_client.o speed.o  s_time.o apps.o s_cb.o s_socket.o app_rand.o version.o 
>sess_id.o  ciphers.o nseq.o pkcs12.o pkcs8.o spkac.o smime.o rand.o engine.o ocsp.o  
>-L.. -lssl  -L.. -lcrypto -Wl,+s -ldld ; \
> > fi
> > /usr/ccs/bin/ld: Unsatisfied symbols:
> >ERR_load_DSO_strings (first referenced in ../libcrypto.a(cryptlib.o)) (code)
> >SHA_Final (first referenced in ../libcrypto.a(cryptlib.o)) (code)
> >ERR_load_EC_strings (first referenced in ../libcrypto.a(cryptlib.o)) (code)
> >lh_delete (first referenced in ../libssl.a(s2_srvr.o)) (code)
> >BIO_snprintf (first referenced in ca.o) (code)
> >BIO_new_socket (first referenced in s_server.o) (code)
> >UI_free (first referenced in apps.o) (code)
> >...
> 
> Danm! The very same thing happened to me earlier! You have to 'find .
> -name lib -size 0 -exec rm {} \;' if you delete libcrypto.a. Sorry:-)
> Alternative is to start from scratch, i.e. unpack into another
> directory, etc. A.

sorry for the delay but i was on christmas holiday.

the problem with the BN_kronecker test seems to occur also with the newly
released 0.9.7 version which seems to include the bn_lcl.h patch you sent
me for testing. i've used fresh source trees to avoid dependency problems
when testing this issue. here're the commands i've used for building:

perl Configure no-asm hpux-parisc2-cc
gmake
gmake test


is anyone using the HP Ansi C compiler for compiling OpenSSL run across
this issue? or am i the only one using this compiler?

-- 
aspahttp://www.kronodoc.fi/



__
OpenSSL Project http://www.openssl.org
Development Mailing List   [EMAIL PROTECTED]
Automated List Manager   [EMAIL PROTECTED]

[openssl.org #410] Re: HP-UX build problems with 0.9.6h

2002-12-30 Thread Andy Polyakov via RT 

I'm resolving this ticket in sincere faith that BN_kronecker issue is a
user environmental problem, i.e. either dependency rules inconsistency
or a compiler bug, and other problems were already addressed.

__
OpenSSL Project http://www.openssl.org
Development Mailing List   [EMAIL PROTECTED]
Automated List Manager   [EMAIL PROTECTED]

Re: [openssl.org #410] Re: HP-UX build problems with 0.9.6h

2002-12-23 Thread Richard Levitte - VMS Whacker via RT 

In message <[EMAIL PROTECTED]> on Mon, 23 Dec 2002 15:14:46 
+0100 (MET), "Andy Polyakov via RT" <[EMAIL PROTECTED]> said:

rt> Danm! The very same thing happened to me earlier! You have to 'find .
rt> -name lib -size 0 -exec rm {} \;' if you delete libcrypto.a. Sorry:-)
rt> Alternative is to start from scratch, i.e. unpack into another
rt> directory, etc. A.

Ahem, may I suggest a 'make clean' as long as things are cleaned up?
:-)

Oh, and the reason libcrypto.a doesn't get rebuilt automatically is
that there are no such targets.  We *should* have them, but that
requires a much larger Makefile cleanup than you'd think...

-- 
Richard Levitte   \ Spannvägen 38, II \ [EMAIL PROTECTED]
Redakteur@Stacken  \ S-168 35  BROMMA  \ T: +46-8-26 52 47
\  SWEDEN   \ or +46-708-26 53 44
Procurator Odiosus Ex Infernis-- [EMAIL PROTECTED]
Member of the OpenSSL development team: http://www.openssl.org/

Unsolicited commercial email is subject to an archival fee of $400.
See  for more info.

__
OpenSSL Project http://www.openssl.org
Development Mailing List   [EMAIL PROTECTED]
Automated List Manager   [EMAIL PROTECTED]

Re: [openssl.org #410] Re: HP-UX build problems with 0.9.6h

2002-12-23 Thread Richard Levitte - VMS Whacker 
In message <[EMAIL PROTECTED]> on Mon, 23 Dec 2002 15:14:46 
+0100 (MET), "Andy Polyakov via RT" <[EMAIL PROTECTED]> said:

rt> Danm! The very same thing happened to me earlier! You have to 'find .
rt> -name lib -size 0 -exec rm {} \;' if you delete libcrypto.a. Sorry:-)
rt> Alternative is to start from scratch, i.e. unpack into another
rt> directory, etc. A.

Ahem, may I suggest a 'make clean' as long as things are cleaned up?
:-)

Oh, and the reason libcrypto.a doesn't get rebuilt automatically is
that there are no such targets.  We *should* have them, but that
requires a much larger Makefile cleanup than you'd think...

-- 
Richard Levitte   \ Spannvägen 38, II \ [EMAIL PROTECTED]
Redakteur@Stacken  \ S-168 35  BROMMA  \ T: +46-8-26 52 47
\  SWEDEN   \ or +46-708-26 53 44
Procurator Odiosus Ex Infernis-- [EMAIL PROTECTED]
Member of the OpenSSL development team: http://www.openssl.org/

Unsolicited commercial email is subject to an archival fee of $400.
See  for more info.
__
OpenSSL Project http://www.openssl.org
Development Mailing List   [EMAIL PROTECTED]
Automated List Manager   [EMAIL PROTECTED]

Re: [openssl.org #410] Re: HP-UX build problems with 0.9.6h

2002-12-23 Thread Andy Polyakov 
> Danm!

Double damn! I can't spell "damn!" Well, it's not like it's appropriate
time for that kind of language:-) A.
__
OpenSSL Project http://www.openssl.org
Development Mailing List   [EMAIL PROTECTED]
Automated List Manager   [EMAIL PROTECTED]

Re: [openssl.org #410] Re: HP-UX build problems with 0.9.6h

2002-12-23 Thread Andy Polyakov via RT 

> i tried this but linking the openssl program as well as the tests fails
> with the following message:
>   cc -o openssl -DMONOLITH -I.. -I../include  -DOPENSSL_THREADS -D_REENTRANT 
>-DDSO_DL -DOPENSSL_NO_KRB5 -DOPENSSL_NO_ASM +DA2.0 +DS2.0 +O3 +Optrs_strongly_typed 
>+Olibcalls -Ae +ESlit -DB_ENDIAN -DMD32_XARRAY openssl.o verify.o asn1pars.o req.o 
>dgst.o dh.o dhparam.o enc.o passwd.o gendh.o errstr.o  ca.o pkcs7.o crl2p7.o crl.o  
>rsa.o rsautl.o dsa.o dsaparam.o  x509.o genrsa.o gendsa.o s_server.o s_client.o 
>speed.o  s_time.o apps.o s_cb.o s_socket.o app_rand.o version.o sess_id.o  ciphers.o 
>nseq.o pkcs12.o pkcs8.o spkac.o smime.o rand.o engine.o ocsp.o  -L.. -lssl  -L.. 
>-lcrypto -Wl,+s -ldld ; \
> fi
> /usr/ccs/bin/ld: Unsatisfied symbols:
>ERR_load_DSO_strings (first referenced in ../libcrypto.a(cryptlib.o)) (code)
>SHA_Final (first referenced in ../libcrypto.a(cryptlib.o)) (code)
>ERR_load_EC_strings (first referenced in ../libcrypto.a(cryptlib.o)) (code)
>lh_delete (first referenced in ../libssl.a(s2_srvr.o)) (code)
>BIO_snprintf (first referenced in ca.o) (code)
>BIO_new_socket (first referenced in s_server.o) (code)
>UI_free (first referenced in apps.o) (code)
>...

Danm! The very same thing happened to me earlier! You have to 'find .
-name lib -size 0 -exec rm {} \;' if you delete libcrypto.a. Sorry:-)
Alternative is to start from scratch, i.e. unpack into another
directory, etc. A.

__
OpenSSL Project http://www.openssl.org
Development Mailing List   [EMAIL PROTECTED]
Automated List Manager   [EMAIL PROTECTED]

Re: [openssl.org #410] Re: HP-UX build problems with 0.9.6h

2002-12-23 Thread Marko Asplund via RT 

On Mon, 23 Dec 2002, Andy Polyakov wrote:

> ...
> Is it actually reproducible? I mean does it end-up in an endless loop
> *every* time you try?

yes, it's reproducible.

> Note that test_kron itself (the procedure printing "test BN_kronecker")
> can't go into an endless loop. It's BN_generate_prime that should take
> the blaim. It's already first line that runs away, isn't it? I mean
> normally you should expect first line being something like "++"
> and then second line with 100 dots exactly.

yes, it is. there are no plusses in the output.

> Now if it's actually reproducible, could you 'rm crypto/bn/*.o
> crypto/bn/lib libcrypto.a', 'make' and confirm that the problem
> persists? The catch is that you also reported problems with 'make
> depend' and your current dependencies might be broken while bn_asm.c is
> not the only module affected by the bug.

i tried this but linking the openssl program as well as the tests fails 
with the following message:
  cc -o openssl -DMONOLITH -I.. -I../include  -DOPENSSL_THREADS -D_REENTRANT 
-DDSO_DL -DOPENSSL_NO_KRB5 -DOPENSSL_NO_ASM +DA2.0 +DS2.0 +O3 +Optrs_strongly_typed 
+Olibcalls -Ae +ESlit -DB_ENDIAN -DMD32_XARRAY openssl.o verify.o asn1pars.o req.o 
dgst.o dh.o dhparam.o enc.o passwd.o gendh.o errstr.o  ca.o pkcs7.o crl2p7.o crl.o  
rsa.o rsautl.o dsa.o dsaparam.o  x509.o genrsa.o gendsa.o s_server.o s_client.o 
speed.o  s_time.o apps.o s_cb.o s_socket.o app_rand.o version.o sess_id.o  ciphers.o 
nseq.o pkcs12.o pkcs8.o spkac.o smime.o rand.o engine.o ocsp.o  -L.. -lssl  -L.. 
-lcrypto -Wl,+s -ldld ; \
fi
/usr/ccs/bin/ld: Unsatisfied symbols:
   ERR_load_DSO_strings (first referenced in ../libcrypto.a(cryptlib.o)) (code)
   SHA_Final (first referenced in ../libcrypto.a(cryptlib.o)) (code)
   ERR_load_EC_strings (first referenced in ../libcrypto.a(cryptlib.o)) (code)
   lh_delete (first referenced in ../libssl.a(s2_srvr.o)) (code)
   BIO_snprintf (first referenced in ca.o) (code)
   BIO_new_socket (first referenced in s_server.o) (code)
   UI_free (first referenced in apps.o) (code)
   ...

best regards,
-- 
aspa
__
OpenSSL Project http://www.openssl.org
Development Mailing List   [EMAIL PROTECTED]
Automated List Manager   [EMAIL PROTECTED]

Re: [openssl.org #410] Re: HP-UX build problems with 0.9.6h

2002-12-23 Thread Andy Polyakov via RT 

> > ...
> > > the warnings and BN test failure seem to be related with the
> > > hpux-parisc2-cc target and no-asm option. without the no-asm option
> > > building goes ok (no warnings, tests succeed) using hpux-parisc2-cc
> > > target.
> >
> > Yet it's extremely important to get it working with no-asm. For that
> > reason I'd like to ask you to double-check that attached patch fixes the
> > problem with no-asm config. A.
> 
> tried it. with your patch i don't get the warnings when compiling bn_asm.c
> and this time BN_sqr test succeeds but BN_kronecker test seems to go into
> an endless loop just printing dots on the screen.

Is it actually reproducible? I mean does it end-up in an endless loop
*every* time you try?

Note that test_kron itself (the procedure printing "test BN_kronecker")
can't go into an endless loop. It's BN_generate_prime that should take
the blaim. It's already first line that runs away, isn't it? I mean
normally you should expect first line being something like "++"
and then second line with 100 dots exactly.

Now if it's actually reproducible, could you 'rm crypto/bn/*.o
crypto/bn/lib libcrypto.a', 'make' and confirm that the problem
persists? The catch is that you also reported problems with 'make
depend' and your current dependencies might be broken while bn_asm.c is
not the only module affected by the bug.

A.

__
OpenSSL Project http://www.openssl.org
Development Mailing List   [EMAIL PROTECTED]
Automated List Manager   [EMAIL PROTECTED]

Re: [openssl.org #410] Re: HP-UX build problems with 0.9.6h

2002-12-23 Thread Marko Asplund via RT 

On Fri, 20 Dec 2002, Andy Polyakov via RT wrote:

> ...
> > the warnings and BN test failure seem to be related with the
> > hpux-parisc2-cc target and no-asm option. without the no-asm option
> > building goes ok (no warnings, tests succeed) using hpux-parisc2-cc
> > target.
> 
> Yet it's extremely important to get it working with no-asm. For that
> reason I'd like to ask you to double-check that attached patch fixes the
> problem with no-asm config. A.

tried it. with your patch i don't get the warnings when compiling bn_asm.c
and this time BN_sqr test succeeds but BN_kronecker test seems to go into
an endless loop just printing dots on the screen.

best regards,
-- 
aspa

__
OpenSSL Project http://www.openssl.org
Development Mailing List   [EMAIL PROTECTED]
Automated List Manager   [EMAIL PROTECTED]

Re: [openssl.org #410] Re: HP-UX build problems with 0.9.6h

2002-12-23 Thread Marko Asplund 
On Fri, 20 Dec 2002, Andy Polyakov via RT wrote:

> ...
> > the warnings and BN test failure seem to be related with the
> > hpux-parisc2-cc target and no-asm option. without the no-asm option
> > building goes ok (no warnings, tests succeed) using hpux-parisc2-cc
> > target.
> 
> Yet it's extremely important to get it working with no-asm. For that
> reason I'd like to ask you to double-check that attached patch fixes the
> problem with no-asm config. A.

tried it. with your patch i don't get the warnings when compiling bn_asm.c
and this time BN_sqr test succeeds but BN_kronecker test seems to go into
an endless loop just printing dots on the screen.

best regards,
-- 
aspa

__
OpenSSL Project http://www.openssl.org
Development Mailing List   [EMAIL PROTECTED]
Automated List Manager   [EMAIL PROTECTED]

Re: [openssl.org #410] Re: HP-UX build problems with 0.9.6h

2002-12-20 Thread Andy Polyakov via RT 

> > > * compilation goes fine apart from the following compiler warning:
> > > cc -I.. -I../.. -I../../include -DOPENSSL_THREADS -D_REENTRANT -DDSO_DL 
>-DOPENSSL_NO_KRB5 -DOPENSSL_NO_IDEA -DOPENSSL_NO_ASM +DA2.0 +DS2.0 +O3 
>+Optrs_strongly_typed +Olibcalls -Ae +ESlit -DB_ENDIAN -DMD32_XARRAY -c bn_asm.c
> > > cc: "bn_asm.c", line 155: warning 950: Shift amount should be in the range of 
>zero through 31 bits.
> > > 
> >
> > I am using HP-UX 10.20 (hpux-parisc-cc target) and have never seen this.
> > I am using 0.9.7-beta? in production.
> >
> > > * tests fail with the following messages:
> > > test BN_lshift
> > > test BN_rshift1
> > > test BN_rshift
> > > test BN_sqr
> > > Square test failed!
> > > *** Error exit code 1
> 
> the warnings and BN test failure seem to be related with the
> hpux-parisc2-cc target and no-asm option. without the no-asm option
> building goes ok (no warnings, tests succeed) using hpux-parisc2-cc
> target.

Yet it's extremely important to get it working with no-asm. For that
reason I'd like to ask you to double-check that attached patch fixes the
problem with no-asm config. A.

--- ./crypto/bn/bn_lcl.h.orig   Mon Dec 16 19:17:24 2002
+++ ./crypto/bn/bn_lcl.hFri Dec 20 18:55:03 2002
@@ -368,7 +368,7 @@
lt=(bl)*(lt); \
m1=(bl)*(ht); \
ht =(bh)*(ht); \
-   m=(m+m1)&BN_MASK2; if (m < m1) ht+=L2HBITS(1L); \
+   m=(m+m1)&BN_MASK2; if (m < m1) ht+=L2HBITS((BN_ULONG)1); \
ht+=HBITS(m); \
m1=L2HBITS(m); \
lt=(lt+m1)&BN_MASK2; if (lt < m1) ht++; \

__
OpenSSL Project http://www.openssl.org
Development Mailing List   [EMAIL PROTECTED]
Automated List Manager   [EMAIL PROTECTED]

Re: [openssl.org #410] Re: HP-UX build problems with 0.9.6h

2002-12-20 Thread Marko Asplund 
On Fri, 20 Dec 2002, Lutz Jaenicke via RT wrote:

> ...
> > * compilation goes fine apart from the following compiler warning:
> > cc -I.. -I../.. -I../../include -DOPENSSL_THREADS -D_REENTRANT -DDSO_DL 
>-DOPENSSL_NO_KRB5 -DOPENSSL_NO_IDEA -DOPENSSL_NO_ASM +DA2.0 +DS2.0 +O3 
>+Optrs_strongly_typed +Olibcalls -Ae +ESlit -DB_ENDIAN -DMD32_XARRAY -c bn_asm.c
> > cc: "bn_asm.c", line 155: warning 950: Shift amount should be in the range of zero 
>through 31 bits.
> > 
> 
> I am using HP-UX 10.20 (hpux-parisc-cc target) and have never seen this.
> I am using 0.9.7-beta? in production.
> 
> > * tests fail with the following messages:
> > test BN_lshift
> > test BN_rshift1
> > test BN_rshift
> > test BN_sqr
> > Square test failed!
> > *** Error exit code 1
> 
> There shouldn't be a problem on 32bit platforms... (PA-RISC2 is able to run
> 64bit, but with the settings provided a 32bit application is built).

the warnings and BN test failure seem to be related with the
hpux-parisc2-cc target and no-asm option. without the no-asm option
building goes ok (no warnings, tests succeed) using hpux-parisc2-cc
target.

best regards,
-- 
aspa


__
OpenSSL Project http://www.openssl.org
Development Mailing List   [EMAIL PROTECTED]
Automated List Manager   [EMAIL PROTECTED]

Re: [openssl.org #410] Re: HP-UX build problems with 0.9.6h

2002-12-20 Thread Marko Asplund 
On Fri, 20 Dec 2002, Richard Levitte - VMS Whacker via RT wrote:

> So that change makes the desired difference?

when building with a target that uses gcc on a system that has makedepend
installed, yes. however, it doesn't fix the original problem when other
compilers are used.

best regards,
-- 
aspa


__
OpenSSL Project http://www.openssl.org
Development Mailing List   [EMAIL PROTECTED]
Automated List Manager   [EMAIL PROTECTED]

Re: [openssl.org #410] Re: HP-UX build problems with 0.9.6h

2002-12-20 Thread Marko Asplund 
On Fri, 20 Dec 2002, Marko Asplund wrote:

> ...
> i tried configuring OpenSSL v0.9.7-beta6 also with the default target
> (hpux-parisc-gcc) using config (with only --prefix argument) but i still
> get the above messages with 'gmake depend'. i think i found the cause for
> this problem:
> ...

i forgot to mention that the patch only fixes this problem when building
with gcc and makedepend is installed on the system. i don't know what
would be the right way to fix this problem when other compilers are used.

best regards,
-- 
aspa
__
OpenSSL Project http://www.openssl.org
Development Mailing List   [EMAIL PROTECTED]
Automated List Manager   [EMAIL PROTECTED]

Re: [openssl.org #410] Re: HP-UX build problems with 0.9.6h

2002-12-20 Thread Marko Asplund 
On Fri, 20 Dec 2002, Lutz Jaenicke via RT wrote:

> ...
> There are several issues in your report, so I have redirected this thread
> into the request tracker.
> 
> > * lots of messages during 'make depend' apparently from the pre-processor:
> > ../util/domd .. -MD makedepend -- -DOPENSSL_THREADS -D_REENTRANT -DDSO_DL 
>-DOPENSSL_NO_KRB5 -DOPENSSL_NO_IDEA -DOPENSSL_NO_ASM +DA2.0 +DS2.0 +O3 
>+Optrs_strongly_typed +Olibcalls -Ae +ESlit -DB_ENDIAN -DMD32_XARRAY -I. -I.. 
>-I../include -DOPENSSL_NO_IDEA  --  cryptlib.c mem.c mem_clr.c mem_dbg.c cversion.c 
>ex_data.c tmdiff.c cpt_err.c ebcdic.c uid.c o_time.c
> > "cryptlib.c":433:  !defined(_POSIX_C_SOURCE) || (_POSIX_C_SOURCE < 199309L)
> 
> Don't know about this one... Do you have "makedepend" on your system,
> otherwise gcc is being used. Maybe some flag is missing.

yes, makedepend is installed on the system (it comes with the HP Ansi C
Developer's Bundle, in the imake package).

i tried configuring OpenSSL v0.9.7-beta6 also with the default target
(hpux-parisc-gcc) using config (with only --prefix argument) but i still
get the above messages with 'gmake depend'. i think i found the cause for
this problem:

--- crypto/Makefile.ssl 2002-12-20 15:01:01.0 +0200
+++ crypto/Makefile.ssl.new 2002-12-20 15:04:47.0 +0200
@@ -141,7 +141,7 @@
@for i in $(SDIRS) ;\
do \
(cd $$i && echo "making depend in crypto/$$i..." && \
-   $(MAKE) MAKEFILE='${MAKEFILE}' INCLUDES='${INCLUDES}' CFLAG='${CFLAG}' 
DEPFLAG='${DEPFLAG}' PERL='${PERL}' depend ); \
+   $(MAKE) MAKEFILE='${MAKEFILE}' INCLUDES='${INCLUDES}' CFLAG='${CFLAG}' 
+DEPFLAG='${DEPFLAG}' MAKEDEPPROG='${MAKEDEPPROG}' PERL='${PERL}' depend ); \
done;
 
 clean:

best regards,
-- 
aspa



__
OpenSSL Project http://www.openssl.org
Development Mailing List   [EMAIL PROTECTED]
Automated List Manager   [EMAIL PROTECTED]

Re: [openssl.org #410] Re: HP-UX build problems with 0.9.6h

2002-12-20 Thread Marko Asplund via RT 


in my problem report email installation problems were mentioned when
OpenSSL 0.9.7-beta6 has been configured to install to a non-default path
e.g. using the something like the following:

perl Configure hpux-parisc2-cc --prefix=/tmp/ossl097b6-6

the attached makefile patch should fix this.

best regards,
-- 
aspa
__
OpenSSL Project http://www.openssl.org
Development Mailing List   [EMAIL PROTECTED]
Automated List Manager   [EMAIL PROTECTED]

Re: [openssl.org #410] Re: HP-UX build problems with 0.9.6h

2002-12-20 Thread Marko Asplund via RT 

On Fri, 20 Dec 2002, Lutz Jaenicke via RT wrote:

> ...
> > * compilation goes fine apart from the following compiler warning:
> > cc -I.. -I../.. -I../../include -DOPENSSL_THREADS -D_REENTRANT -DDSO_DL 
>-DOPENSSL_NO_KRB5 -DOPENSSL_NO_IDEA -DOPENSSL_NO_ASM +DA2.0 +DS2.0 +O3 
>+Optrs_strongly_typed +Olibcalls -Ae +ESlit -DB_ENDIAN -DMD32_XARRAY -c bn_asm.c
> > cc: "bn_asm.c", line 155: warning 950: Shift amount should be in the range of zero 
>through 31 bits.
> > 
> 
> I am using HP-UX 10.20 (hpux-parisc-cc target) and have never seen this.
> I am using 0.9.7-beta? in production.
> 
> > * tests fail with the following messages:
> > test BN_lshift
> > test BN_rshift1
> > test BN_rshift
> > test BN_sqr
> > Square test failed!
> > *** Error exit code 1
> 
> There shouldn't be a problem on 32bit platforms... (PA-RISC2 is able to run
> 64bit, but with the settings provided a 32bit application is built).

the warnings and BN test failure seem to be related with the
hpux-parisc2-cc target and no-asm option. without the no-asm option
building goes ok (no warnings, tests succeed) using hpux-parisc2-cc
target.

best regards,
-- 
aspa



__
OpenSSL Project http://www.openssl.org
Development Mailing List   [EMAIL PROTECTED]
Automated List Manager   [EMAIL PROTECTED]

Re: [openssl.org #410] Re: HP-UX build problems with 0.9.6h

2002-12-20 Thread Marko Asplund via RT 

On Fri, 20 Dec 2002, Richard Levitte - VMS Whacker via RT wrote:

> So that change makes the desired difference?

when building with a target that uses gcc on a system that has makedepend
installed, yes. however, it doesn't fix the original problem when other
compilers are used.

best regards,
-- 
aspa



__
OpenSSL Project http://www.openssl.org
Development Mailing List   [EMAIL PROTECTED]
Automated List Manager   [EMAIL PROTECTED]

Re: [openssl.org #410] Re: HP-UX build problems with 0.9.6h

2002-12-20 Thread Richard Levitte - VMS Whacker via RT 

So that change makes the desired difference?

In message <[EMAIL PROTECTED]> on Fri, 20 Dec 2002 14:12:50 
+0100 (MET), "Marko Asplund via RT" <[EMAIL PROTECTED]> said:

rt> i tried configuring OpenSSL v0.9.7-beta6 also with the default target
rt> (hpux-parisc-gcc) using config (with only --prefix argument) but i still
rt> get the above messages with 'gmake depend'. i think i found the cause for
rt> this problem:
rt> 
rt> --- crypto/Makefile.ssl 2002-12-20 15:01:01.0 +0200
rt> +++ crypto/Makefile.ssl.new 2002-12-20 15:04:47.0 +0200
rt> @@ -141,7 +141,7 @@
rt> @for i in $(SDIRS) ;\
rt> do \
rt> (cd $$i && echo "making depend in crypto/$$i..." && \
rt> -   $(MAKE) MAKEFILE='${MAKEFILE}' INCLUDES='${INCLUDES}' CFLAG='${CFLAG}' 
DEPFLAG='${DEPFLAG}' PERL='${PERL}' depend ); \
rt> +   $(MAKE) MAKEFILE='${MAKEFILE}' INCLUDES='${INCLUDES}' CFLAG='${CFLAG}' 
DEPFLAG='${DEPFLAG}' MAKEDEPPROG='${MAKEDEPPROG}' PERL='${PERL}' depend ); \
rt> done;

-- 
Richard Levitte   \ Spannvägen 38, II \ [EMAIL PROTECTED]
Redakteur@Stacken  \ S-168 35  BROMMA  \ T: +46-8-26 52 47
\  SWEDEN   \ or +46-708-26 53 44
Procurator Odiosus Ex Infernis-- [EMAIL PROTECTED]
Member of the OpenSSL development team: http://www.openssl.org/

Unsolicited commercial email is subject to an archival fee of $400.
See  for more info.

__
OpenSSL Project http://www.openssl.org
Development Mailing List   [EMAIL PROTECTED]
Automated List Manager   [EMAIL PROTECTED]

Re: [openssl.org #410] Re: HP-UX build problems with 0.9.6h

2002-12-20 Thread Richard Levitte - VMS Whacker 
So that change makes the desired difference?

In message <[EMAIL PROTECTED]> on Fri, 20 Dec 2002 14:12:50 
+0100 (MET), "Marko Asplund via RT" <[EMAIL PROTECTED]> said:

rt> i tried configuring OpenSSL v0.9.7-beta6 also with the default target
rt> (hpux-parisc-gcc) using config (with only --prefix argument) but i still
rt> get the above messages with 'gmake depend'. i think i found the cause for
rt> this problem:
rt> 
rt> --- crypto/Makefile.ssl 2002-12-20 15:01:01.0 +0200
rt> +++ crypto/Makefile.ssl.new 2002-12-20 15:04:47.0 +0200
rt> @@ -141,7 +141,7 @@
rt> @for i in $(SDIRS) ;\
rt> do \
rt> (cd $$i && echo "making depend in crypto/$$i..." && \
rt> -   $(MAKE) MAKEFILE='${MAKEFILE}' INCLUDES='${INCLUDES}' CFLAG='${CFLAG}' 
DEPFLAG='${DEPFLAG}' PERL='${PERL}' depend ); \
rt> +   $(MAKE) MAKEFILE='${MAKEFILE}' INCLUDES='${INCLUDES}' CFLAG='${CFLAG}' 
DEPFLAG='${DEPFLAG}' MAKEDEPPROG='${MAKEDEPPROG}' PERL='${PERL}' depend ); \
rt> done;

-- 
Richard Levitte   \ Spannvägen 38, II \ [EMAIL PROTECTED]
Redakteur@Stacken  \ S-168 35  BROMMA  \ T: +46-8-26 52 47
\  SWEDEN   \ or +46-708-26 53 44
Procurator Odiosus Ex Infernis-- [EMAIL PROTECTED]
Member of the OpenSSL development team: http://www.openssl.org/

Unsolicited commercial email is subject to an archival fee of $400.
See  for more info.
__
OpenSSL Project http://www.openssl.org
Development Mailing List   [EMAIL PROTECTED]
Automated List Manager   [EMAIL PROTECTED]

Re: [openssl.org #410] Re: HP-UX build problems with 0.9.6h

2002-12-20 Thread Marko Asplund via RT 

On Fri, 20 Dec 2002, Marko Asplund wrote:

> ...
> i tried configuring OpenSSL v0.9.7-beta6 also with the default target
> (hpux-parisc-gcc) using config (with only --prefix argument) but i still
> get the above messages with 'gmake depend'. i think i found the cause for
> this problem:
> ...

i forgot to mention that the patch only fixes this problem when building
with gcc and makedepend is installed on the system. i don't know what
would be the right way to fix this problem when other compilers are used.

best regards,
-- 
aspa

__
OpenSSL Project http://www.openssl.org
Development Mailing List   [EMAIL PROTECTED]
Automated List Manager   [EMAIL PROTECTED]

Re: [openssl.org #410] Re: HP-UX build problems with 0.9.6h

2002-12-20 Thread Marko Asplund via RT 

On Fri, 20 Dec 2002, Lutz Jaenicke via RT wrote:

> ...
> There are several issues in your report, so I have redirected this thread
> into the request tracker.
> 
> > * lots of messages during 'make depend' apparently from the pre-processor:
> > ../util/domd .. -MD makedepend -- -DOPENSSL_THREADS -D_REENTRANT -DDSO_DL 
>-DOPENSSL_NO_KRB5 -DOPENSSL_NO_IDEA -DOPENSSL_NO_ASM +DA2.0 +DS2.0 +O3 
>+Optrs_strongly_typed +Olibcalls -Ae +ESlit -DB_ENDIAN -DMD32_XARRAY -I. -I.. 
>-I../include -DOPENSSL_NO_IDEA  --  cryptlib.c mem.c mem_clr.c mem_dbg.c cversion.c 
>ex_data.c tmdiff.c cpt_err.c ebcdic.c uid.c o_time.c
> > "cryptlib.c":433:  !defined(_POSIX_C_SOURCE) || (_POSIX_C_SOURCE < 199309L)
> 
> Don't know about this one... Do you have "makedepend" on your system,
> otherwise gcc is being used. Maybe some flag is missing.

yes, makedepend is installed on the system (it comes with the HP Ansi C
Developer's Bundle, in the imake package).

i tried configuring OpenSSL v0.9.7-beta6 also with the default target
(hpux-parisc-gcc) using config (with only --prefix argument) but i still
get the above messages with 'gmake depend'. i think i found the cause for
this problem:

--- crypto/Makefile.ssl 2002-12-20 15:01:01.0 +0200
+++ crypto/Makefile.ssl.new 2002-12-20 15:04:47.0 +0200
@@ -141,7 +141,7 @@
@for i in $(SDIRS) ;\
do \
(cd $$i && echo "making depend in crypto/$$i..." && \
-   $(MAKE) MAKEFILE='${MAKEFILE}' INCLUDES='${INCLUDES}' CFLAG='${CFLAG}' 
DEPFLAG='${DEPFLAG}' PERL='${PERL}' depend ); \
+   $(MAKE) MAKEFILE='${MAKEFILE}' INCLUDES='${INCLUDES}' CFLAG='${CFLAG}' 
+DEPFLAG='${DEPFLAG}' MAKEDEPPROG='${MAKEDEPPROG}' PERL='${PERL}' depend ); \
done;
 
 clean:

best regards,
-- 
aspa




__
OpenSSL Project http://www.openssl.org
Development Mailing List   [EMAIL PROTECTED]
Automated List Manager   [EMAIL PROTECTED]

Re: [openssl.org #410] Re: HP-UX build problems with 0.9.6h

2002-12-20 Thread Richard Levitte - VMS Whacker via RT 

In message <[EMAIL PROTECTED]> on Fri, 20 Dec 2002 09:01:40 
+0100 (MET), "Lutz Jaenicke via RT" <[EMAIL PROTECTED]> said:

rt> > * lots of messages during 'make depend' apparently from the pre-processor:
rt> > ../util/domd .. -MD makedepend -- -DOPENSSL_THREADS -D_REENTRANT -DDSO_DL 
-DOPENSSL_NO_KRB5 -DOPENSSL_NO_IDEA -DOPENSSL_NO_ASM +DA2.0 +DS2.0 +O3 
+Optrs_strongly_typed +Olibcalls -Ae +ESlit -DB_ENDIAN -DMD32_XARRAY -I. -I.. 
-I../include -DOPENSSL_NO_IDEA  --  cryptlib.c mem.c mem_clr.c mem_dbg.c cversion.c 
ex_data.c tmdiff.c cpt_err.c ebcdic.c uid.c o_time.c
rt> > "cryptlib.c":433:  !defined(_POSIX_C_SOURCE) || (_POSIX_C_SOURCE < 199309L)
rt> 
rt> Don't know about this one... Do you have "makedepend" on your system,
rt> otherwise gcc is being used. Maybe some flag is missing.

Small correction: gcc is used instead of makedepend only if the
compiler is exactly "gcc".  The following line in Configure sets that
up:

s/^MAKEDEPPROG=.*$/MAKEDEPPROG= $cc/ if $cc eq "gcc";

-- 
Richard Levitte   \ Spannvägen 38, II \ [EMAIL PROTECTED]
Redakteur@Stacken  \ S-168 35  BROMMA  \ T: +46-8-26 52 47
\  SWEDEN   \ or +46-708-26 53 44
Procurator Odiosus Ex Infernis-- [EMAIL PROTECTED]
Member of the OpenSSL development team: http://www.openssl.org/

Unsolicited commercial email is subject to an archival fee of $400.
See  for more info.

__
OpenSSL Project http://www.openssl.org
Development Mailing List   [EMAIL PROTECTED]
Automated List Manager   [EMAIL PROTECTED]

Re: [openssl.org #410] Re: HP-UX build problems with 0.9.6h

2002-12-20 Thread Richard Levitte - VMS Whacker 
In message <[EMAIL PROTECTED]> on Fri, 20 Dec 2002 09:01:40 
+0100 (MET), "Lutz Jaenicke via RT" <[EMAIL PROTECTED]> said:

rt> > * lots of messages during 'make depend' apparently from the pre-processor:
rt> > ../util/domd .. -MD makedepend -- -DOPENSSL_THREADS -D_REENTRANT -DDSO_DL 
-DOPENSSL_NO_KRB5 -DOPENSSL_NO_IDEA -DOPENSSL_NO_ASM +DA2.0 +DS2.0 +O3 
+Optrs_strongly_typed +Olibcalls -Ae +ESlit -DB_ENDIAN -DMD32_XARRAY -I. -I.. 
-I../include -DOPENSSL_NO_IDEA  --  cryptlib.c mem.c mem_clr.c mem_dbg.c cversion.c 
ex_data.c tmdiff.c cpt_err.c ebcdic.c uid.c o_time.c
rt> > "cryptlib.c":433:  !defined(_POSIX_C_SOURCE) || (_POSIX_C_SOURCE < 199309L)
rt> 
rt> Don't know about this one... Do you have "makedepend" on your system,
rt> otherwise gcc is being used. Maybe some flag is missing.

Small correction: gcc is used instead of makedepend only if the
compiler is exactly "gcc".  The following line in Configure sets that
up:

s/^MAKEDEPPROG=.*$/MAKEDEPPROG= $cc/ if $cc eq "gcc";

-- 
Richard Levitte   \ Spannvägen 38, II \ [EMAIL PROTECTED]
Redakteur@Stacken  \ S-168 35  BROMMA  \ T: +46-8-26 52 47
\  SWEDEN   \ or +46-708-26 53 44
Procurator Odiosus Ex Infernis-- [EMAIL PROTECTED]
Member of the OpenSSL development team: http://www.openssl.org/

Unsolicited commercial email is subject to an archival fee of $400.
See  for more info.
__
OpenSSL Project http://www.openssl.org
Development Mailing List   [EMAIL PROTECTED]
Automated List Manager   [EMAIL PROTECTED]

Re: [openssl.org #410] Re: HP-UX build problems with 0.9.6h

2002-12-20 Thread Lutz Jaenicke via RT 

On Fri, Dec 20, 2002 at 08:48:57AM +0100, Marko Asplund via RT wrote:
> 
> On Wed, 18 Dec 2002, Lutz Jaenicke wrote:
> 
> > ...
> > What about 0.9.7-beta6? The shared library build system was reworked for
> > 0.9.7 (with a special treatment to handle HP-UX).
> 
> ok, tried building 0.9.7-beta6 using the same configuration and procedure.
> here are some notes about the build process:

There are several issues in your report, so I have redirected this thread
into the request tracker.

> * lots of messages during 'make depend' apparently from the pre-processor:
> ../util/domd .. -MD makedepend -- -DOPENSSL_THREADS -D_REENTRANT -DDSO_DL 
>-DOPENSSL_NO_KRB5 -DOPENSSL_NO_IDEA -DOPENSSL_NO_ASM +DA2.0 +DS2.0 +O3 
>+Optrs_strongly_typed +Olibcalls -Ae +ESlit -DB_ENDIAN -DMD32_XARRAY -I. -I.. 
>-I../include -DOPENSSL_NO_IDEA  --  cryptlib.c mem.c mem_clr.c mem_dbg.c cversion.c 
>ex_data.c tmdiff.c cpt_err.c ebcdic.c uid.c o_time.c
> "cryptlib.c":433:  !defined(_POSIX_C_SOURCE) || (_POSIX_C_SOURCE < 199309L)

Don't know about this one... Do you have "makedepend" on your system,
otherwise gcc is being used. Maybe some flag is missing.

> * compilation goes fine apart from the following compiler warning:
> cc -I.. -I../.. -I../../include -DOPENSSL_THREADS -D_REENTRANT -DDSO_DL 
>-DOPENSSL_NO_KRB5 -DOPENSSL_NO_IDEA -DOPENSSL_NO_ASM +DA2.0 +DS2.0 +O3 
>+Optrs_strongly_typed +Olibcalls -Ae +ESlit -DB_ENDIAN -DMD32_XARRAY -c bn_asm.c
> cc: "bn_asm.c", line 155: warning 950: Shift amount should be in the range of zero 
>through 31 bits.
> 

I am using HP-UX 10.20 (hpux-parisc-cc target) and have never seen this.
I am using 0.9.7-beta? in production.

> * tests fail with the following messages:
> test BN_lshift
> test BN_rshift1
> test BN_rshift
> test BN_sqr
> Square test failed!
> *** Error exit code 1

There shouldn't be a problem on 32bit platforms... (PA-RISC2 is able to run
64bit, but with the settings provided a 32bit application is built).

Best regards,
Lutz
-- 
Lutz Jaenicke [EMAIL PROTECTED]
http://www.aet.TU-Cottbus.DE/personen/jaenicke/
BTU Cottbus, Allgemeine Elektrotechnik
Universitaetsplatz 3-4, D-03044 Cottbus

__
OpenSSL Project http://www.openssl.org
Development Mailing List   [EMAIL PROTECTED]
Automated List Manager   [EMAIL PROTECTED]

[openssl.org #410] Re: HP-UX build problems with 0.9.6h

2002-12-19 Thread Marko Asplund via RT 

On Wed, 18 Dec 2002, Lutz Jaenicke wrote:

> ...
> What about 0.9.7-beta6? The shared library build system was reworked for
> 0.9.7 (with a special treatment to handle HP-UX).

ok, tried building 0.9.7-beta6 using the same configuration and procedure.
here are some notes about the build process:

* lots of messages during 'make depend' apparently from the pre-processor:
../util/domd .. -MD makedepend -- -DOPENSSL_THREADS -D_REENTRANT -DDSO_DL 
-DOPENSSL_NO_KRB5 -DOPENSSL_NO_IDEA -DOPENSSL_NO_ASM +DA2.0 +DS2.0 +O3 
+Optrs_strongly_typed +Olibcalls -Ae +ESlit -DB_ENDIAN -DMD32_XARRAY -I. -I.. 
-I../include -DOPENSSL_NO_IDEA  --  cryptlib.c mem.c mem_clr.c mem_dbg.c cversion.c 
ex_data.c tmdiff.c cpt_err.c ebcdic.c uid.c o_time.c
"cryptlib.c":433:  !defined(_POSIX_C_SOURCE) || (_POSIX_C_SOURCE < 199309L)
* compilation goes fine apart from the following compiler warning:
cc -I.. -I../.. -I../../include -DOPENSSL_THREADS -D_REENTRANT -DDSO_DL 
-DOPENSSL_NO_KRB5 -DOPENSSL_NO_IDEA -DOPENSSL_NO_ASM +DA2.0 +DS2.0 +O3 
+Optrs_strongly_typed +Olibcalls -Ae +ESlit -DB_ENDIAN -DMD32_XARRAY -c bn_asm.c
cc: "bn_asm.c", line 155: warning 950: Shift amount should be in the range of zero 
through 31 bits.

* tests fail with the following messages:
test BN_lshift
test BN_rshift1
test BN_rshift
test BN_sqr
Square test failed!
*** Error exit code 1

Stop.
* installation fails

i'm attaching a full typescript in this email.

best regards,
-- 
aspa

__
OpenSSL Project http://www.openssl.org
Development Mailing List   [EMAIL PROTECTED]
Automated List Manager   [EMAIL PROTECTED]