From: open...@master.openssl.org (OpenSSL)
> OpenSSL version 1.0.0 Beta 4
> [...]
> Since the third beta, the following has happened:
> [...]
- Build system fixes including VMS.
> [...]
Not entirely successful. Around here:
ALP $ cc /version
HP C V7.3-009 on OpenVMS Alpha V8.3
I tried:
@ makevms.com ALL "" NODEBUG DECC TCPIP
@ [.test]tests.com
ALP $ gdiff -u makevms.com_orig makevms.com
--- makevms.com_orig 2009-08-25 02:30:02 -0500
+++ makevms.com 2009-11-11 13:21:47 -0600
@@ -349,7 +349,7 @@
$!
$! There are many places where this is needed.
$!
-$ WRITE H_FILE "#define _XOPEN_SOURCE_EXTENDED"
+$!!! WRITE H_FILE "#define _XOPEN_SOURCE_EXTENDED"
$!
$! Close the [.CRYPTO.<ARCH>]OPENSSLCONF.H file
$!
What, too many things were working correctly? This _seriously_
breaks the build. Why was this added? "is needed" is not a helpful
explanation.
--- apps/install.com_orig 2009-05-15 11:37:04 -0500
+++ apps/install.com 2009-11-11 14:16:15 -0600
@@ -57,7 +57,7 @@
$
$ SET NOON
$ COPY CA.COM WRK_SSLEXE:CA.COM/LOG
-$ SET FILE/PROT=W:RE WRK_SSLVEXE:CA.COM
+$ SET FILE/PROT=W:RE WRK_SSLEXE:CA.COM
$ COPY OPENSSL-VMS.CNF WRK_SSLROOT:[000000]OPENSSL.CNF/LOG
$ SET FILE/PROT=W:R WRK_SSLROOT:[000000]OPENSSL.CNF
$ SET ON
Beside being simpler and perhaps a bit faster, using COPY /PROTECTION
instead of separate COPY and SET FILE /PROTECTION commands (as
previously suggested) would halve the opportunities for careless errors
of this type.
--- crypto/crypto-lib.com_orig 2009-08-25 02:22:08 -0500
+++ crypto/crypto-lib.com 2009-11-11 10:48:40 -0600
@@ -193,7 +193,8 @@
$ LIB_SEED = "seed,seed_ecb,seed_cbc,seed_cfb,seed_ofb"
$ LIB_MODES = "cbc128,ctr128,cfb128,ofb128"
$ LIB_BN_ASM = "[.asm]vms.mar,vms-helper"
-$ IF F$TRNLNM("OPENSSL_NO_ASM") THEN LIB_BN_ASM = "bn_asm"
+$ IF F$TRNLNM("OPENSSL_NO_ASM") .OR. ARCH .NES. "VAX" THEN -
+ LIB_BN_ASM = "bn_asm"
$ LIB_BN = "bn_add,bn_div,bn_exp,bn_lib,bn_ctx,bn_mul,bn_mod,"+ -
"bn_print,bn_rand,bn_shift,bn_word,bn_blind,"+ -
"bn_kron,bn_sqrt,bn_gcd,bn_prime,bn_err,bn_sqr,"+LIB_BN_ASM+","+ -
Even if MACRO32 code were faster on an Alpha, the MACRO32 compiler
there won't compile "vms.mar".
--- crypto/symhacks.h_orig 2009-05-15 11:00:08 -0500
+++ crypto/symhacks.h 2009-11-11 10:56:52 -0600
@@ -138,6 +138,8 @@
#define X509_policy_node_get0_qualifiers X509_pcy_node_get0_qualifiers
#undef X509_STORE_CTX_get_explicit_policy
#define X509_STORE_CTX_get_explicit_policy X509_STORE_CTX_get_expl_policy
+#undef X509_STORE_CTX_get0_current_issuer
+#define X509_STORE_CTX_get0_current_issuer X509_STORE_CTX_get0_current_iss
/* Hack some long CRYPTO names */
#undef CRYPTO_set_dynlock_destroy_callback
Yet another %CC-W-LONGEXTERN complaint.
--- util/libeay.num_orig 2009-11-04 07:29:58 -0600
+++ util/libeay.num 2009-11-11 14:00:31 -0600
@@ -4168,4 +4168,5 @@
X509_STORE_set_verify_cb 4543 EXIST::FUNCTION:
X509_STORE_CTX_get0_current_crl 4544 EXIST::FUNCTION:
X509_STORE_CTX_get0_parent_ctx 4545 EXIST::FUNCTION:
-X509_STORE_CTX_get0_current_issuer 4546 EXIST::FUNCTION:
+X509_STORE_CTX_get0_current_issuer 4546 EXIST:!VMS:FUNCTION:
+X509_STORE_CTX_get0_current_iss 4546 EXIST:VMS:FUNCTION:
See "crypto/symhacks.h".
"test/testenc.com" seems to fail. SSLROOT not defined? (If you
thought that it worked, what were you testing?) Apparently,
"test/tests.com" exits on error, so no test results after that.
Are there any plans to get this stuff to work properly before the
actual release? The beta kits so far have not been encouraging. I've
given up on seeing several previously suggested changed adopted, but it
would be nice if, for example, a simple build simply worked.
------------------------------------------------------------------------
Steven M. Schweda s...@antinode-info
382 South Warwick Street (+1) 651-699-9818
Saint Paul MN 55105-2547
______________________________________________________________________
OpenSSL Project http://www.openssl.org
Development Mailing List openssl-dev@openssl.org
Automated List Manager majord...@openssl.org
