Re: pod2html target for Windows port
On 2003-03-01 10:10, Ulf Möller wrote: This patch adds a `html' target to the Windows build, that uses pod2html to make html documentation; e.g. There is nothing Visual C specific to the patch, or is it? Not that I know of, but as I said, it's only been tested with MSVC, since that's all I have. A potential issue would be the makefile macros $, $(D) and $(B). -- Peter 'Luna' Runestig (fd. Altberg), Sweden [EMAIL PROTECTED] PGP Key ID: 0xD07BBE13 Fingerprint: 7B5C 1F48 2997 C061 DE4B 42EA CB99 A35C D07B BE13 AOL Instant Messenger Screen name: PRunestig __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
pod2html target for Windows port
Hi all! This patch adds a `html' target to the Windows build, that uses pod2html to make html documentation; e.g. nmake -f ms\nt.mak html It has only been tested with MSVC6 and native perl (i.e. not Cygwin's version). diff -burN openssl-0.9.7a.orig/ms/ms-point.pl openssl-0.9.7a/ms/ms-point.pl --- openssl-0.9.7a.orig/ms/ms-point.pl 1970-01-01 01:00:00.0 +0100 +++ openssl-0.9.7a/ms/ms-point.pl 2003-02-27 15:33:10.0 +0100 @@ -0,0 +1,15 @@ +# A Microsoft Windows platform version of util/point.sh, adapted to work around +# the native command shell's limitations. + +($#ARGV == 1) || die usage: ms-point.pl directory basename\n; +$dir = $ARGV[0]; +$dir =~ s/\\/\//g; # in case this is used with non-native perl +$base = $ARGV[1]; + +while (STDIN) { +/(\S+)/; +if ($1 and $1 ne $base) { + link($dir/$base.html, $dir/$1.html); + printf($1.html = $base.html\n); +} +} diff -burN openssl-0.9.7a.orig/util/mk1mf.pl openssl-0.9.7a/util/mk1mf.pl --- openssl-0.9.7a.orig/util/mk1mf.pl 2003-02-14 06:20:28.0 +0100 +++ openssl-0.9.7a/util/mk1mf.pl2003-02-27 15:30:53.0 +0100 @@ -631,6 +631,26 @@ $rules.=do_link_rule(\$(BIN_D)$o\$(E_EXE)$exep,\$(E_OBJ),\$(LIBS_DEP),\$(L_LIBS) \$(EX_LIBS)); +# add rules and targets for .pod - .html file generation +if (($platform eq VC-WIN32) || ($platform eq VC-NT)) { +$rules.=EOF; +.SUFFIXES: .pod + +.pod.html: +# we must convert e.g. SSL_CTX_new(3) - SSL_CTX_new in the pod file + [EMAIL PROTECTED] -pi~ -e s/\\([0123456789]\\)//g \$ + pod2html --htmlroot=.. --podroot=doc --podpath=apps:crypto:ssl --header --infile=\$ --outfile=\$*.html + [EMAIL PROTECTED] util/extract-names.pl \$ | perl ms/ms-point.pl \$(D) \$(B) + +html: \$(PODS:.pod=.html) +EOF +$defs.=PODS=; +foreach (glob(doc/apps/*.pod)) { s/\//\\/g; $defs.=$_ ; } +foreach (glob(doc/crypto/*.pod)) { s/\//\\/g; $defs.=$_ ; } +foreach (glob(doc/ssl/*.pod)) { s/\//\\/g; $defs.=$_ ; } +$defs.=\n\n; +} + print $defs; if ($platform eq linux-elf) { Cheers, - Peter -- Peter 'Luna' Runestig (fd. Altberg), Sweden [EMAIL PROTECTED] PGP Key ID: 0xD07BBE13 Fingerprint: 7B5C 1F48 2997 C061 DE4B 42EA CB99 A35C D07B BE13 AOL Instant Messenger Screen name: PRunestig __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: OpenSSL and compression using ZLIB
Gregory Stark wrote: Oops, I meant 2246. And reading it more carefully, I agree with your interpretation. The dictionary need not be reset. Compression state can and should be maintained across records. So, is anyone working on improving the zlib code according to these new guidelines? Cheers, - Peter -- Peter 'Luna' Runestig (fd. Altberg), Sweden [EMAIL PROTECTED] PGP Key ID: 0xD07BBE13 Fingerprint: 7B5C 1F48 2997 C061 DE4B 42EA CB99 A35C D07B BE13 AOL Instant Messenger Screen name: PRunestig __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
SSL_CTX_set_cipher_list(ALL) negotiates ADH-AES256-SHA
Hi all, I have a client and a server, that sets SSL_CTX_set_cipher_list(ALL) (and SSL_CTX_set_tmp_dh_callback()). With beta3 and beta4, that makes the negotiated cipher to be ADH-AES256-SHA. I would expect something like DHE-RSA-AES256-SHA instead (which I get if I do SSL_CTX_set_cipher_list(ALL:!ADH) instead), or am I missing something? Isn't the strongest common cipher expected to be used? Doing it all with 0.9.6g negotiates EXP1024-RC4-SHA instead, which is also a surprise to me; e.g. SSL_CTX_set_cipher_list(MEDIUM) gives IDEA-CBC-SHA. Cheers, - Peter -- Peter 'Luna' Runestig (fd. Altberg), Sweden [EMAIL PROTECTED] PGP Key ID: 0xD07BBE13 Fingerprint: 7B5C 1F48 2997 C061 DE4B 42EA CB99 A35C D07B BE13 AOL Instant Messenger Screen name: PRunestig __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: Windows Installer MSI creation code, take two
ftp://ftp.runestig.com/pub/openssl-win32/openssl-0.9.6g-pod2html-20020912.patch http://www.runestig.com/pub/openssl-win32/openssl-0.9.6g-pod2html-20020912.patch This one adds a 'html' target to the makefiles (for Windows), that creates .html documentation files from the .pod files, using perl utility 'pod2html'. After playing around with, and understanding a little bit more, 'pod2html', here is a much better patch: ftp://ftp.runestig.com/pub/openssl-win32/openssl-0.9.6g-pod2html-20020917.patch http://www.runestig.com/pub/openssl-win32/openssl-0.9.6g-pod2html-20020917.patch Cheers, - Peter -- Peter 'Luna' Runestig (fd. Altberg), Sweden [EMAIL PROTECTED] PGP Key ID: 0xD07BBE13 Fingerprint: 7B5C 1F48 2997 C061 DE4B 42EA CB99 A35C D07B BE13 AOL Instant Messenger Screenname: PRunestig __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Win32: Windows Installer MSI package creation code
Hi! Just FYI: Mostly as an exercise for myself, I'm working on a toolkit to create Windows Installer MSI packages from C source code. I'm thinking that it might be useful in Win32 open source projects that wants to roll a MSI package during the standard make process. There is still a lot of functions to add to the toolkit, but the base functionality is there. I decided to test the toolkit so far on OpenSSL, since it's basically just packing a bunch of files (and adding to PATH). If anyone is curious, here is a patch: ftp://ftp.runestig.com/pub/misc/openssl-0.9.6g-msi-20020828.zip http://www.runestig.com/pub/misc/openssl-0.9.6g-msi-20020828.zip It might look big, since there's some bitmaps and icons represented in a header file in there. After applying the patch, and doing the Windows Configure stuff, you can do something like this to build and get a MSI package: nmake -f 32dll.mak msi Cheers, - Peter Peter 'Luna' Runestig (fd. Altberg), Sweden [EMAIL PROTECTED] PGP Key ID: 0xD07BBE13 Fingerprint: 7B5C 1F48 2997 C061 DE4B 42EA CB99 A35C D07B BE13 AOL Instant Messenger Screenname: PRunestig __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Win32: Add version info to the dll
Hi all! Attached is a patch (against 0.9.6g) that adds version info to libeay32.dll and ssleay32.dll, for the Win32 build using MSVC. It's probably a bit rought, since it's all modifications to the perl configuration code, and I'm not very familiar with perl. And I'm not the best person to decide e.g. what copyright info to include. But it seems to work OK, and I think version info in those dlls can be useful. Cheers, - Peter Peter 'Luna' Runestig (fd. Altberg), Sweden [EMAIL PROTECTED] PGP Key ID: 0xD07BBE13 Fingerprint: 7B5C 1F48 2997 C061 DE4B 42EA CB99 A35C D07B BE13 AOL Instant Messenger Screenname: PRunestig openssl-0.9.6g-dllversion.patch.gz Description: GNU Zip compressed data
#ifdef DEBUG printf() in crypto/rand/rand_win.c (and others)
Hi all, I just found a bunch of these things in crypto/rand/rand_win.c, that surprised me: #ifdef DEBUG printf(randomness from PROV_RSA_FULL\n); #endif I'm using a standard debug build on Windows, and was, as I said, surprised by the extra info on screen when running my app. Maybe they should a) be changed to e.g. #ifdef DEBUG_INFO or something else not default defined. b) at least use stderr, fprintf(stderr, randomness from PROV_RSA_FULL\n) Cheers, - Peter Peter 'Luna' Runestig (fd. Altberg), Sweden PGP Key ID: 0xD07BBE13 Fingerprint: 7B5C 1F48 2997 C061 DE4B 42EA CB99 A35C D07B BE13 AOL Instant Messenger Screenname: PRunestig __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: Please try today's snapshot
From: "Richard Levitte - VMS Whacker" [EMAIL PROTECTED]: I'd like those who have reported problems with OpenSSL 0.9.6abeta1 to try tonights snapshots: openssl-0.9.6-stable-SNAP-20010314.tar.gz openssl-engine-0.9.6-stable-SNAP-20010314.tar.gz Well, it seems I have managed to confuse myself a bit. According to my notes, I built openssl-0.9.6, on Win32 using VC 6, like this: perl Configure VC-WIN32 ms\32all.bat But that doesn't work. At the time, I must have executed "ms\do_ms.bat" before I did "ms\32all.bat". And "perl Configure VC-WIN32" isn't needed, it's done inside "ms\32all.bat". So, in order to build openssl-0.9.6-stable-SNAP-20010314, this additional change is needed: --- ms\32all.bat.orig Tue Mar 13 23:03:28 2001 +++ ms\32all.batWed Mar 14 21:36:46 2001 @@ -6,6 +6,8 @@ perl util\mk1mf.pl %OPTS% VC-WIN32 32.mak perl util\mk1mf.pl %OPTS% debug dll VC-WIN32 d32dll.mak perl util\mk1mf.pl %OPTS% dll VC-WIN32 32dll.mak +perl util\mkdef.pl 32 libeay ms\libeay32.def +perl util\mkdef.pl 32 ssleay ms\ssleay32.def nmake -f d32.mak nmake -f 32.mak But with that change, all four flavours of openssl is successfully built (and tested using "ms\test.bat") on Windows 2000, using VC 6.0, in a single stroke using "ms\32all.bat". Sorry about the confusion. - Peter ---- Peter 'Luna' Runestig (fd. Altberg), Sweden [EMAIL PROTECTED] PGP Key ID: 0xD07BBE13 Fingerprint: 7B5C 1F48 2997 C061 DE4B 42EA CB99 A35C D07B BE13 AOL Instant Messenger Screenname: PRunestig __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: [ANNOUNCEMENT] OpenSSL 0.9.6a Beta 1 released
It seems like a line is missing in 'ms\32all.bat': set OPTS=no-asm perl Configure VC-WIN32 perl util\mkfiles.pl MINFO NEW perl util\mk1mf.pl %OPTS% debug VC-WIN32 d32.mak perl util\mk1mf.pl %OPTS% VC-WIN32 32.mak perl util\mk1mf.pl %OPTS% debug dll VC-WIN32 d32dll.mak perl util\mk1mf.pl %OPTS% dll VC-WIN32 32dll.mak nmake -f d32.mak nmake -f 32.mak nmake -f d32dll.mak nmake -f 32dll.mak Peter 'Luna' Runestig (fd. Altberg), Sweden [EMAIL PROTECTED] PGP Key ID: 0xD07BBE13 Fingerprint: 7B5C 1F48 2997 C061 DE4B 42EA CB99 A35C D07B BE13 AOL Instant Messenger Screenname: PRunestig __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Problems openssl-SNAP-20001219 and Win32
Hi all, These changes where necessary to build openssl static (dll link failes anyway, but I don't have time to try to fix that) with VC++ 6.0, W2K, options "nasm -DNO_KRB5" and using ms\32all.bat. Note that there's some type casting to get rid of warings about converting `long´ to something smaller size; maybe some funamental change is needed, I don't know. I'm sure the core folks won't just blindly commit this... :-) (Lines will be wrapped, but maybe it's preffered over an attached file?) diff -ur openssl-SNAP-20001219.orig/apps/engine.c openssl-SNAP-20001219/apps/engine.c --- openssl-SNAP-20001219.orig/apps/engine.c Tue Nov 07 00:00:18 2000 +++ openssl-SNAP-20001219/apps/engine.c Wed Dec 20 09:18:14 2000 @@ -83,7 +83,7 @@ return; } -static int append_buf(char **buf, char *s, int *size, int step) +static int append_buf(char **buf, char *s, unsigned int *size, int step) { int l = strlen(s); @@ -203,7 +203,7 @@ } if (list_cap) { -int cap_size = 256; +unsigned int cap_size = 256; char *cap_buf = NULL; if (ENGINE_get_RSA(e) != NULL diff -ur openssl-SNAP-20001219.orig/crypto/asn1/tasn_dec.c openssl-SNAP-20001219/crypto/asn1/tasn_dec.c --- openssl-SNAP-20001219.orig/crypto/asn1/tasn_dec.c Fri Dec 08 20:07:06 2000 +++ openssl-SNAP-20001219/crypto/asn1/tasn_dec.c Wed Dec 20 08:12:18 2000 @@ -58,6 +58,9 @@ #include stddef.h +#ifdef WIN32 +#include memory.h +#endif /* WIN32 */ #include openssl/asn1.h #include openssl/asn1t.h #include openssl/objects.h @@ -201,7 +204,7 @@ if(tag != -1) { p = *in; imphack = *p; - *p = (*p V_ASN1_CONSTRUCTED) | it-utype; + *p = (unsigned char)((*p V_ASN1_CONSTRUCTED) | it-utype); } ptmpval = cf-asn1_d2i(pval, in, len); @@ -315,7 +318,7 @@ * it increases efficiency in some cases. */ if(i == (it-tcount - 1)) isopt = 0; - else isopt = seqtt-flags ASN1_TFLG_OPTIONAL; + else isopt = (char)(seqtt-flags ASN1_TFLG_OPTIONAL); /* attempt to read in field, allowing each to be OPTIONAL */ ret = asn1_template_ex_d2i(pseqval, p, len, seqtt, isopt, ctx); if(!ret) { diff -ur openssl-SNAP-20001219.orig/crypto/asn1/tasn_enc.c openssl-SNAP-20001219/crypto/asn1/tasn_enc.c --- openssl-SNAP-20001219.orig/crypto/asn1/tasn_enc.c Thu Dec 14 01:00:18 2000 +++ openssl-SNAP-20001219/crypto/asn1/tasn_enc.c Wed Dec 20 08:08:26 2000 @@ -58,6 +58,9 @@ #include stddef.h +#ifdef WIN32 +#include memory.h +#endif /* WIN32 */ #include openssl/asn1.h #include openssl/asn1t.h #include openssl/objects.h diff -ur openssl-SNAP-20001219.orig/crypto/asn1/tasn_new.c openssl-SNAP-20001219/crypto/asn1/tasn_new.c --- openssl-SNAP-20001219.orig/crypto/asn1/tasn_new.c Fri Dec 08 20:07:06 2000 +++ openssl-SNAP-20001219/crypto/asn1/tasn_new.c Wed Dec 20 08:07:51 2000 @@ -58,6 +58,9 @@ #include stddef.h +#ifdef WIN32 +#include memory.h +#endif /* WIN32 */ #include openssl/asn1.h #include openssl/objects.h #include openssl/err.h diff -ur openssl-SNAP-20001219.orig/crypto/asn1/tasn_utl.c openssl-SNAP-20001219/crypto/asn1/tasn_utl.c --- openssl-SNAP-20001219.orig/crypto/asn1/tasn_utl.c Fri Dec 08 20:07:08 2000 +++ openssl-SNAP-20001219/crypto/asn1/tasn_utl.c Wed Dec 20 08:12:58 2000 @@ -58,6 +58,9 @@ #include stddef.h +#ifdef WIN32 +#include memory.h +#endif /* WIN32 */ #include openssl/asn1.h #include openssl/asn1t.h #include openssl/objects.h diff -ur openssl-SNAP-20001219.orig/crypto/objects/o_names.c openssl-SNAP-20001219/crypto/objects/o_names.c --- openssl-SNAP-20001219.orig/crypto/objects/o_names.c Wed Dec 13 19:00:34 2000 +++ openssl-SNAP-20001219/crypto/objects/o_names.c Wed Dec 20 08:04:10 2000 @@ -295,12 +295,12 @@ d.n=0; OBJ_NAME_do_all(type,do_all_sorted_fn,d); - qsort(d.names,d.n,sizeof *d.names,do_all_sorted_cmp); + qsort((void *)d.names,d.n,sizeof *d.names,do_all_sorted_cmp); for(n=0 ; n d.n ; ++n) fn(d.names[n],arg); - OPENSSL_free(d.names); + OPENSSL_free((void *)d.names); } static int free_type; -------- Peter "Luna" Runestig (fd. Altberg), Sweden [EMAIL PROTECTED] PGP Key ID: 0xD07BBE13 Fingerprint: 7B5C 1F48 2997 C061 DE4B 42EA CB99 A35C D07B BE13 AOL Instant Messenger Screenname: PRunestig __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Proposed change to ms\32all.bat
... so the build stops at first error, and don't repeat it four times. diff -ur openssl-SNAP-20001219.orig/ms/32all.bat openssl-SNAP-20001219/ms/32all.bat --- openssl-SNAP-20001219.orig/ms/32all.bat Sat Jan 30 13:00:14 1999 +++ openssl-SNAP-20001219/ms/32all.bat Wed Dec 20 12:28:25 2000 @@ -7,6 +7,11 @@ perl util\mk1mf.pl %OPTS% dll VC-WIN32 32dll.mak nmake -f d32.mak +@if errorlevel 1 goto end nmake -f 32.mak +@if errorlevel 1 goto end nmake -f d32dll.mak +@if errorlevel 1 goto end nmake -f 32dll.mak + +:end Peter "Luna" Runestig (fd. Altberg), Sweden [EMAIL PROTECTED] PGP Key ID: 0xD07BBE13 Fingerprint: 7B5C 1F48 2997 C061 DE4B 42EA CB99 A35C D07B BE13 AOL Instant Messenger Screenname: PRunestig __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Patch: X509_NAME_print()
Hi all, When you do a $ openssl x509 -text cert.pem the format of the subject and issuer lines is a mix of "old" and "new" styles, like this: Subject: C=SE, L=Bollnaes, CN=Peter 'Luna' [EMAIL PROTECTED]/UID=luna This happens for every "object short name" that is longer than two chars. Here is a patch to make it print like this: Subject: C=SE, L=Bollnaes, CN=Peter 'Luna' Runestig, [EMAIL PROTECTED], UID=luna diff -ur openssl-SNAP-20001219.orig/crypto/asn1/t_x509.c openssl-SNAP-20001219/crypto/asn1/t_x509.c --- openssl-SNAP-20001219.orig/crypto/asn1/t_x509.c Thu Dec 14 20:01:06 2000 +++ openssl-SNAP-20001219/crypto/asn1/t_x509.c Wed Dec 20 13:48:40 2000 @@ -429,6 +429,27 @@ return(0); } +static int is_obj_name(const char *s) +{ +#define MAX_OBJ_NAME 5 + char *ss, *eq = strchr(s, '='); + if (eq == NULL) + return 0; + if ((eq - s) MAX_OBJ_NAME) + return 0; + for (ss = s; ss eq; ss++) { +#ifndef CHARSET_EBCDIC + if (!( ((*ss = 'A') (*ss = 'Z')) || + ((*ss = 'a') (*ss = 'z')) )) + return 0; +#else + if (!( isupper(*ss) || islower(*ss) )) + return 0; +#endif /* CHARSET_EBCDIC */ + } + return 1; +} + int X509_NAME_print(BIO *bp, X509_NAME *name, int obase) { char *s,*c; @@ -444,23 +465,7 @@ c=s; for (;;) { -#ifndef CHARSET_EBCDIC - if (((*s == '/') - ((s[1] = 'A') (s[1] = 'Z') ( - (s[2] == '=') || - ((s[2] = 'A') (s[2] = 'Z') - (s[3] == '=')) -))) || - (*s == '\0')) -#else - if (((*s == '/') - (isupper(s[1]) ( - (s[2] == '=') || - (isupper(s[2]) - (s[3] == '=')) -))) || - (*s == '\0')) -#endif + if ( ((*s == '/') is_obj_name(s+1)) || (*s == '\0')) { if ((l = 0) !first) { -- Peter 'Luna' Runestig (fd. Altberg), Sweden [EMAIL PROTECTED] PGP Key ID: 0xD07BBE13 Fingerprint: 7B5C 1F48 2997 C061 DE4B 42EA CB99 A35C D07B BE13 AOL Instant Messenger Screenname: PRunestig __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Status of (zlib) compression?
Hi all, I just found out that openssl includes a zlib compression implementation, if you configure it with -DZLIB. It seems to work fine, except that the comp/decomp functions in crypto/comp/c_zlib.c has fprintf(stderr, ...) debug output in them. Maybe those should be dropped, or at least wrapped with some option or something? Or is there some reason one should not use this (yet)? Cheers, Peter -- Peter "Luna" Runestig (fd. Altberg), Sweden [EMAIL PROTECTED] PGP Key ID: 0xD07BBE13 Fingerprint: 7B5C 1F48 2997 C061 DE4B 42EA CB99 A35C D07B BE13 Gubben Movitz ler och nickar, / men från Charons mörka sund dödens blund / i dina blickar / bådar snart din sista stund. Carl Michael Bellman, Fredmans epistel nr 34 __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
6 more beta3 successes (1 small issue though)
Built and tested ok: Linux i386 2.2.17 / glibc 2.1.3 / gcc 2.95.2 OpenSSL 0.9.6-beta3 21 Sep 2000 built on: Thu Sep 21 17:01:08 CEST 2000 platform: linux-elf options: bn(64,32) md2(int) rc4(idx,int) des(ptr,risc1,16,long) idea(int) blowfish(idx) compiler: gcc -fPIC -DTHREADS -D_REENTRANT -DDSO_DLFCN -DHAVE_DLFCN_H -DL_ENDIAN -DTERMIO -O3 -fomit-frame-pointer -m486 -Wall -DSHA1_ASM -DMD5_ASM -DRMD160_ASM Linux i386 2.0.36 / glibc 2.0.7 / gcc 2.7.2.3 - OpenSSL 0.9.6-beta3 21 Sep 2000 built on: Thu Sep 21 17:21:54 CEST 2000 platform: linux-elf options: bn(64,32) md2(int) rc4(idx,int) des(ptr,risc1,16,long) idea(int) blowfish(idx) compiler: gcc -fPIC -DTHREADS -D_REENTRANT -DDSO_DLFCN -DHAVE_DLFCN_H -DL_ENDIAN -DTERMIO -O3 -fomit-frame-pointer -m486 -Wall -DSHA1_ASM -DMD5_ASM -DRMD160_ASM OpenBSD 2.7 GENERIC#25 i386 / gcc 2.95.2 OpenSSL 0.9.6-beta3 21 Sep 2000 built on: Thu Sep 21 18:15:09 CEST 2000 platform: OpenBSD-x86 options: bn(64,32) md2(int) rc4(idx,int) des(ptr,risc1,16,long) idea(int) blowfish(idx) compiler: gcc -DDSO_DLFCN -DHAVE_DLFCN_H -DL_ENDIAN -DTERMIOS -O3 -fomit-frame-pointer -m486 -DSHA1_ASM -DMD5_ASM -DRMD160_ASM Compaq Tru64 OSF1 V4.0 1229 alpha - OpenSSL 0.9.6-beta3 21 Sep 2000 built on: Thu Sep 21 12:58:27 EDT 2000 platform: alpha-cc options: bn(64,64) md2(int) rc4(ptr,int) des(idx,cisc,4,long) idea(int) blowfish(idx) compiler: cc -DDSO_DLFCN -DHAVE_DLFCN_H -std1 -tune host -O4 -readonly_strings Solaris 8 Intel (SunOS 5.8 Generic_108529-01 i86pc i386 i86pc) / gcc 2.95.2 --- OpenSSL 0.9.6-beta3 21 Sep 2000 built on: Thu Sep 21 18:51:14 MEST 2000 platform: solaris-x86-gcc options: bn(64,32) md2(int) rc4(idx,int) des(ptr,risc1,16,long) idea(int) blowfish(idx) compiler: gcc -fPIC -DTHREADS -D_REENTRANT -DDSO_DLFCN -DHAVE_DLFCN_H -O3 -fomit-frame-pointer -m486 -Wall -DL_ENDIAN -DNO_INLINE_ASM -DSHA1_ASM -DMD5_ASM -DRMD160_ASM FreeBSD 4.1-RELEASE alpha / gcc 2.95.2 -- (with Richard's one-line patch) OpenSSL 0.9.6-beta3 21 Sep 2000 built on: Thu Sep 21 13:12:25 EDT 2000 platform: FreeBSD-alpha options: bn(64,64) md2(int) rc4(ptr,int) des(ptr,risc2,4,int) idea(int) blowfish(idx) compiler: gcc -DTERMIOS -O3 -fomit-frame-pointer *BUT*: I get these warnings: gcc -I. -I../include -DTERMIOS -O3 -fomit-frame-pointer -c mem_dbg.c cc1: warning: *** *** The -O3 flag TRIGGERS KNOWN OPTIMIZER BUGS ON THIS PLATFORM *** -- Peter 'Luna' Runestig (fd. Altberg), Sweden [EMAIL PROTECTED] PGP Key ID: 0xD07BBE13 Fingerprint: 7B5C 1F48 2997 C061 DE4B 42EA CB99 A35C D07B BE13 AOL Instant Messenger Screenname: PRunestig __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: 6 more beta3 successes (1 small issue though)
Richard Levitte - VMS Whacker wrote: From: "Peter 'Luna' Runestig" [EMAIL PROTECTED] peter+openssl-dev gcc -I. -I../include -DTERMIOS -O3 -fomit-frame-pointer -c mem_dbg.c peter+openssl-dev cc1: warning: peter+openssl-dev *** peter+openssl-dev *** The -O3 flag TRIGGERS KNOWN OPTIMIZER BUGS ON THIS PLATFORM peter+openssl-dev *** Do you know if it should be lower of higher? [ "Never assume", said a wise woman once ] No, I'm afraid not. I don't have that much experience with alpha. -- Peter 'Luna' Runestig (fd. Altberg), Sweden [EMAIL PROTECTED] PGP Key ID: 0xD07BBE13 Fingerprint: 7B5C 1F48 2997 C061 DE4B 42EA CB99 A35C D07B BE13 AOL Instant Messenger Screenname: PRunestig __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: Win-32 Crash Report (beta2)
Isn't this contributing by a US citizen? ;-) -- Peter 'Luna' Runestig (fd. Altberg), Sweden [EMAIL PROTECTED] PGP Key ID: 0xD07BBE13 Fingerprint: 7B5C 1F48 2997 C061 DE4B 42EA CB99 A35C D07B BE13 AOL Instant Messenger Screenname: PRunestig __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
First error compiling openssl-0.9.5beta1 on NT 4
Hi all! I have just successfully built openssl-0.9.5beta1 on NT 4 with VC++ 6.0 and MASM 6.14. I had to "tweak" two thing to get it through. My setup is this: perl Configure VC-WIN32 ms\do_masm nmake -f ms\ntdll.mak I'm not on this list, so you may CC any comments to me. First one, short version I need this patch to make openssl-0.9.5beta1 compile clean with MASM 6: --- crypto/perlasm/x86ms.pl.origMon Jul 12 20:12:40 1999 +++ crypto/perlasm/x86ms.pl Fri Feb 25 10:27:27 2000 @@ -341,7 +341,7 @@ $label{$_[0]}="${label}${_[0]}"; $label++; } - push(@out,"$label{$_[0]}:\n"); + push(@out,"$label{$_[0]}::\n"); } sub main'data_word Why? I get this error: ml /Cp /coff /c /Cx /Focrypto\sha\asm\s1-win32.obj .\crypto\sha\asm\s1-win32.asm Microsoft (R) Macro Assembler Version 6.14.8444 Copyright (C) Microsoft Corp 1981-1997. All rights reserved. Assembling: .\crypto\sha\asm\s1-win32.asm .\crypto\sha\asm\s1-win32.asm(1711) : error A2006: undefined symbol : L001shortcut NMAKE : fatal error U1077: 'ml' : return code '0x1' Stop. This is (hopefully) the interesting parts of crypto\sha\asm\s1-win32.asm: TITLE sha1-586.asm .486 .model FLAT _TEXT SEGMENT PUBLIC _sha1_block_asm_data_order _sha1_block_asm_data_order PROC NEAR snip mov DWORD PTR 132[esp],esi L001shortcut: ; snip ret _TEXT SEGMENT PUBLIC _sha1_block_asm_host_order _sha1_block_asm_host_order PROC NEAR snip jmp L001shortcut _sha1_block_asm_host_order ENDP _TEXT ENDS _sha1_block_asm_data_order ENDP _TEXT ENDS END Now, I'm by no means a MASM expert, but it seems that labels with a single colon, ":", is scope dependent. So the "L001shortcut:" label isn't visible in the scope of "_sha1_block_asm_host_order PROC NEAR". To make labels "global", they should be suffixed with double colons, "::", like "L001shortcut::". Cheers, - Peter Peter "Luna" Runestig (fd. Altberg), Sweden [EMAIL PROTECTED] PGP Key ID: 0x33BE83E0 Fingerprint: 05DB BFF0 4F9D 1FFA 8441 E859 7F7B E52E 33BE 83E0 Gubben Movitz ler och nickar, / men från Charons mörka sund dödens blund / i dina blickar / bådar snart din sista stund. Carl Michael Bellman, Fredmans epistel nr 34 __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Second error compiling openssl-0.9.5beta1 on NT 4
Hi all! I have just successfully built openssl-0.9.5beta1 on NT 4 with VC++ 6.0 and MASM 6.14. I had to "tweak" two thing to get it through. My setup is this: perl Configure VC-WIN32 ms\do_masm nmake -f ms\ntdll.mak I'm not on this list, so you may CC any comments to me. Second one, short version = This patch also needed: --- ssl/s3_pkt.c.orig Tue Feb 22 03:59:24 2000 +++ ssl/s3_pkt.cFri Feb 25 10:39:17 2000 @@ -846,7 +846,7 @@ if (dest_maxlen 0) { n = dest_maxlen - *dest_len; /* available space in 'dest' */ - if (rr-length n) + if (rr-length (unsigned int)n) n = rr-length; /* available bytes */ /* now move 'n' bytes: */ Why? I get this error, which speaks for itself: cl /Fotmp32dll\s3_pkt.obj -Iinc32 -Itmp32dll /MD /W3 /WX /G5 /Ox /O2 /Ob2 /Gs0 /GF /Gy /nologo -DWIN32 -DWIN32_LEAN_AND_MEAN -DL_ENDIAN -DBN_ASM -DMD5_ASM -DSHA1_A SM -DRMD160_ASM /Fdout32dll /GD -D_WINDLL -D_DLL -c .\ssl\s3_pkt.c s3_pkt.c .\ssl\s3_pkt.c(849) : error C2220: warning treated as error - no object file generated .\ssl\s3_pkt.c(849) : warning C4018: '' : signed/unsigned mismatch NMAKE : fatal error U1077: 'cl' : return code '0x2' Stop. Cheers, - Peter -------- Peter "Luna" Runestig (fd. Altberg), Sweden [EMAIL PROTECTED] PGP Key ID: 0x33BE83E0 Fingerprint: 05DB BFF0 4F9D 1FFA 8441 E859 7F7B E52E 33BE 83E0 Gubben Movitz ler och nickar, / men från Charons mörka sund dödens blund / i dina blickar / bådar snart din sista stund. Carl Michael Bellman, Fredmans epistel nr 34 __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
First error compiling openssl-0.9.5beta1 on NT 4
Hi all! (sorry about posting to the wrong list, openssl-dev) I have just successfully built openssl-0.9.5beta1 on NT 4 with VC++ 6.0 and MASM 6.14. I had to "tweak" two thing to get it through. My setup is this: perl Configure VC-WIN32 ms\do_masm nmake -f ms\ntdll.mak First one, short version I need this patch to make openssl-0.9.5beta1 compile clean with MASM 6: --- crypto/perlasm/x86ms.pl.origMon Jul 12 20:12:40 1999 +++ crypto/perlasm/x86ms.pl Fri Feb 25 10:27:27 2000 @@ -341,7 +341,7 @@ $label{$_[0]}="${label}${_[0]}"; $label++; } - push(@out,"$label{$_[0]}:\n"); + push(@out,"$label{$_[0]}::\n"); } sub main'data_word Why? I get this error: ml /Cp /coff /c /Cx /Focrypto\sha\asm\s1-win32.obj .\crypto\sha\asm\s1-win32.asm Microsoft (R) Macro Assembler Version 6.14.8444 Copyright (C) Microsoft Corp 1981-1997. All rights reserved. Assembling: .\crypto\sha\asm\s1-win32.asm .\crypto\sha\asm\s1-win32.asm(1711) : error A2006: undefined symbol : L001shortcut NMAKE : fatal error U1077: 'ml' : return code '0x1' Stop. This is (hopefully) the interesting parts of crypto\sha\asm\s1-win32.asm: TITLE sha1-586.asm .486 .model FLAT _TEXT SEGMENT PUBLIC _sha1_block_asm_data_order _sha1_block_asm_data_order PROC NEAR snip mov DWORD PTR 132[esp],esi L001shortcut: ; snip ret _TEXT SEGMENT PUBLIC _sha1_block_asm_host_order _sha1_block_asm_host_order PROC NEAR snip jmp L001shortcut _sha1_block_asm_host_order ENDP _TEXT ENDS _sha1_block_asm_data_order ENDP _TEXT ENDS END Now, I'm by no means a MASM expert, but it seems that labels with a single colon, ":", is scope dependent. So the "L001shortcut:" label isn't visible in the scope of "_sha1_block_asm_host_order PROC NEAR". To make labels "global", they should be suffixed with double colons, "::", like "L001shortcut::". Cheers, - Peter Peter "Luna" Runestig (fd. Altberg), Sweden [EMAIL PROTECTED] PGP Key ID: 0x33BE83E0 Fingerprint: 05DB BFF0 4F9D 1FFA 8441 E859 7F7B E52E 33BE 83E0 Gubben Movitz ler och nickar, / men från Charons mörka sund dödens blund / i dina blickar / bådar snart din sista stund. Carl Michael Bellman, Fredmans epistel nr 34 __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Second error compiling openssl-0.9.5beta1 on NT 4
Hi all! (sorry about posting to the wrong list, openssl-dev) I have just successfully built openssl-0.9.5beta1 on NT 4 with VC++ 6.0 and MASM 6.14. I had to "tweak" two thing to get it through. My setup is this: perl Configure VC-WIN32 ms\do_masm nmake -f ms\ntdll.mak Second one, short version = This patch also needed: --- ssl/s3_pkt.c.orig Tue Feb 22 03:59:24 2000 +++ ssl/s3_pkt.cFri Feb 25 10:39:17 2000 @@ -846,7 +846,7 @@ if (dest_maxlen 0) { n = dest_maxlen - *dest_len; /* available space in 'dest' */ - if (rr-length n) + if (rr-length (unsigned int)n) n = rr-length; /* available bytes */ /* now move 'n' bytes: */ Why? I get this error, which speaks for itself: cl /Fotmp32dll\s3_pkt.obj -Iinc32 -Itmp32dll /MD /W3 /WX /G5 /Ox /O2 /Ob2 /Gs0 /GF /Gy /nologo -DWIN32 -DWIN32_LEAN_AND_MEAN -DL_ENDIAN -DBN_ASM -DMD5_ASM -DSHA1_A SM -DRMD160_ASM /Fdout32dll /GD -D_WINDLL -D_DLL -c .\ssl\s3_pkt.c s3_pkt.c .\ssl\s3_pkt.c(849) : error C2220: warning treated as error - no object file generated .\ssl\s3_pkt.c(849) : warning C4018: '' : signed/unsigned mismatch NMAKE : fatal error U1077: 'cl' : return code '0x2' Stop. Cheers, - Peter -------- Peter "Luna" Runestig (fd. Altberg), Sweden [EMAIL PROTECTED] PGP Key ID: 0x33BE83E0 Fingerprint: 05DB BFF0 4F9D 1FFA 8441 E859 7F7B E52E 33BE 83E0 Gubben Movitz ler och nickar, / men från Charons mörka sund dödens blund / i dina blickar / bådar snart din sista stund. Carl Michael Bellman, Fredmans epistel nr 34 __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Bug report with patch
Hi, Problem: If the negotiated cipher is ADH (ie, the SSL_aNULL flag is set) and if the verify mode is SSL_VERIFY_PEER, the server will send a certificate request to the client. The receipt of this request by the client is considered a fatal protocol error in TLS. Therefore, the request should not be sent. Fix: The following patch to s3_srvr.c prevents the sending of the certificate request by the server when the cipher suite is anonymous. *** backup\s3_srvr.c Tue Nov 16 19:00:34 1999 --- s3_srvr.c Sat Dec 18 14:55:40 1999 *** *** 287,293 case SSL3_ST_SW_CERT_REQ_A: case SSL3_ST_SW_CERT_REQ_B: ! if (!(s-verify_mode SSL_VERIFY_PEER) || ((s-session-peer != NULL) (s-verify_mode SSL_VERIFY_CLIENT_ONCE))) { --- 287,294 case SSL3_ST_SW_CERT_REQ_A: case SSL3_ST_SW_CERT_REQ_B: ! if ((s-s3-tmp.new_cipher-algorithms SSL_aNULL) || ! !(s-verify_mode SSL_VERIFY_PEER) || ((s-session-peer != NULL) (s-verify_mode SSL_VERIFY_CLIENT_ONCE))) { Cheers, - Peter __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]