oweals/openssl.git
25 years agoLet ssl_get_prev_session reliably work in multi-threaded settings.
Bodo Möller [Sun, 23 May 1999 13:07:03 +0000 (13:07 +0000)]
Let ssl_get_prev_session reliably work in multi-threaded settings.

25 years agoSet ext_list to NULL after use.
Dr. Stephen Henson [Sun, 23 May 1999 12:02:18 +0000 (12:02 +0000)]
Set ext_list to NULL after use.

25 years agoAvoid memory hole when we don't like the session proposed by the client
Bodo Möller [Sun, 23 May 1999 10:43:46 +0000 (10:43 +0000)]
Avoid memory hole when we don't like the session proposed by the client

25 years agoFixup do_nt.bat for new mk1mf arg order.
Dr. Stephen Henson [Sat, 22 May 1999 23:24:38 +0000 (23:24 +0000)]
Fixup do_nt.bat for new mk1mf arg order.

25 years agoMove the Win32 #undefs of X509_NAME and PKCS7_ISSUER_AND_SERIAL so they will
Dr. Stephen Henson [Fri, 21 May 1999 12:14:35 +0000 (12:14 +0000)]
Move the Win32 #undefs of X509_NAME and PKCS7_ISSUER_AND_SERIAL so they will
always get included with the relevant files.

25 years agoWe need e_os.h here.
Bodo Möller [Fri, 21 May 1999 11:46:29 +0000 (11:46 +0000)]
We need e_os.h here.

25 years agoRemove traces of bad idea :-)
Bodo Möller [Fri, 21 May 1999 11:21:19 +0000 (11:21 +0000)]
Remove traces of bad idea :-)

25 years agoOlder versions of SINIX C development system inherited a bug from
Bodo Möller [Fri, 21 May 1999 11:20:04 +0000 (11:20 +0000)]
Older versions of SINIX C development system inherited a bug from
SGI' cc; enable workaround.
Submitted by: Martin Kraemer

25 years agoIt was a very bad idea to use #include "../e_os.h" -- when this occurs
Bodo Möller [Fri, 21 May 1999 11:16:48 +0000 (11:16 +0000)]
It was a very bad idea to use #include "../e_os.h" -- when this occurs
in cryptlib.h (which is often included as "../cryptlib.h"), then the
question remains relative to which directory this is to be interpreted.
gcc went one further directory up, as intended; but makedepend thinks
differently, and so probably do some C compilers.  So the ../ must go away;
thus e_os.h goes back into include/openssl (but I now use
#include "openssl/e_os.h" instead of <openssl/e_os.h> to make the point) --
and we have another huge bunch of dependency changes.  Argh.

25 years agoDon't include top-level CVS directory.
Ben Laurie [Fri, 21 May 1999 08:38:17 +0000 (08:38 +0000)]
Don't include top-level CVS directory.

25 years agoThe last argument in the d2i_XXX_fp and d2i_XXX_bio functions should be
Dr. Stephen Henson [Fri, 21 May 1999 01:06:23 +0000 (01:06 +0000)]
The last argument in the d2i_XXX_fp and  d2i_XXX_bio functions should be
of type XXX ** not XXX *

25 years agoAdd a kludge :-(
Bodo Möller [Thu, 20 May 1999 22:18:49 +0000 (22:18 +0000)]
Add a kludge :-(
There were problems with putting e_os.h just into the top directory,
because the test programs are compiled within test/ in the "standard"
case in in their original directories in the makefile.one case;
and in the latter symlinks may not be available.

25 years agoDon't install e_os.h in include/openssl, use it only as a local
Bodo Möller [Thu, 20 May 1999 21:59:20 +0000 (21:59 +0000)]
Don't install e_os.h in include/openssl, use it only as a local
include file.

25 years agoBring VMS in sync with the recent changes.
Ulf Möller [Thu, 20 May 1999 21:00:29 +0000 (21:00 +0000)]
Bring VMS in sync with the recent changes.

Submitted by: Richard Levitte <levitte@stacken.kth.se>

25 years agoOn seconds thoughts, the version number shoud _never_ decrease.
Ben Laurie [Thu, 20 May 1999 19:57:53 +0000 (19:57 +0000)]
On seconds thoughts, the version number shoud _never_ decrease.

25 years agoRevert.
Ben Laurie [Thu, 20 May 1999 19:46:23 +0000 (19:46 +0000)]
Revert.

25 years agoPrepare for a beta release.
Ben Laurie [Thu, 20 May 1999 19:33:46 +0000 (19:33 +0000)]
Prepare for a beta release.

25 years agoNeXT doesn't have dirent.
Ulf Möller [Thu, 20 May 1999 17:58:42 +0000 (17:58 +0000)]
NeXT doesn't have dirent.

Pointed out by Juergen Moellenhoff <jurgen@oic.de>

25 years agoDetect Siemens platforms.
Ulf Möller [Thu, 20 May 1999 17:28:19 +0000 (17:28 +0000)]
Detect Siemens platforms.

Submitted by: Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>

25 years agoSome solaris-usparc MD5 fixes.
Bodo Möller [Thu, 20 May 1999 13:11:43 +0000 (13:11 +0000)]
Some solaris-usparc MD5 fixes.

25 years agoAvoid a warning.
Ulf Möller [Thu, 20 May 1999 02:09:46 +0000 (02:09 +0000)]
Avoid a warning.

Submitted by: Sylvain Robitaille <syl@alcor.concordia.ca>

25 years agoDeclare test key data as static.
Ulf Möller [Thu, 20 May 1999 01:46:39 +0000 (01:46 +0000)]
Declare test key data as static.

Submitted by: Andy Polyakov <appro@fy.chalmers.se>

25 years agoChange self-description (mklink.pl is not just faster than the last
Bodo Möller [Thu, 20 May 1999 01:43:28 +0000 (01:43 +0000)]
Change self-description (mklink.pl is not just faster than the last
released version of mklink.sh -- the latter couldn't even handle
the ../../include/openssl case).

25 years agoBignum library bug fix. IRIX 6 passes "make test" now!
Ulf Möller [Thu, 20 May 1999 01:43:07 +0000 (01:43 +0000)]
Bignum library bug fix. IRIX 6 passes "make test" now!
This also avoids the problems with SC4.2 and unpatched SC5.

Submitted by: Andy Polyakov <appro@fy.chalmers.se>

25 years agoInclude some notes on basic extension usage and change openssl.cnf to usually
Dr. Stephen Henson [Wed, 19 May 1999 23:54:58 +0000 (23:54 +0000)]
Include some notes on basic extension usage and change openssl.cnf to usually
do sensible things with extensions.

25 years agoAdd -mv8 to solaris-usparc-oldgcc.
Bodo Möller [Wed, 19 May 1999 22:45:08 +0000 (22:45 +0000)]
Add -mv8 to solaris-usparc-oldgcc.

25 years agoTypos.
Ulf Möller [Wed, 19 May 1999 19:20:49 +0000 (19:20 +0000)]
Typos.

25 years agoOops.
Ulf Möller [Wed, 19 May 1999 19:14:59 +0000 (19:14 +0000)]
Oops.

25 years agoSmall corrections.
Ulf Möller [Wed, 19 May 1999 18:39:23 +0000 (18:39 +0000)]
Small corrections.

25 years agoNote that the numbering scheme used to be different.
Bodo Möller [Wed, 19 May 1999 18:08:35 +0000 (18:08 +0000)]
Note that the numbering scheme used to be different.

25 years agoSwitch to new version numbering scheme.
Ben Laurie [Wed, 19 May 1999 17:36:40 +0000 (17:36 +0000)]
Switch to new version numbering scheme.

25 years agoConvert gcc version detection (for solaris-usparc-gcc) into a form
Bodo Möller [Wed, 19 May 1999 16:50:16 +0000 (16:50 +0000)]
Convert gcc version detection (for solaris-usparc-gcc) into a form
better suited for finding what went wrong in case that some compiler
versions create an output we can't parse.

25 years agoBugfix: GCCVAR contains two lines ("Reading specs ..." and the actual
Bodo Möller [Wed, 19 May 1999 16:35:21 +0000 (16:35 +0000)]
Bugfix: GCCVAR contains two lines ("Reading specs ..." and the actual
version), so we need
   echo $GCCVAR | sed ...
instead of
   echo "$GCCVAR" | sed ...
to process it as intended.

25 years agoNew functions sk_set, sk_value and sk_num to replace existing macros: this is
Dr. Stephen Henson [Wed, 19 May 1999 12:45:16 +0000 (12:45 +0000)]
New functions sk_set, sk_value and sk_num to replace existing macros: this is
to minimise the effects on existing code.

25 years agoFor Solaris with old gcc versions (that don't understand
Bodo Möller [Wed, 19 May 1999 12:27:33 +0000 (12:27 +0000)]
For Solaris with old gcc versions (that don't understand
-mcpu=ultrasparc), don't silently step to solaris-sparc-gcc, which
would unnecessarily use less efficient assembler code.
Instead, use new configuration variant solaris-usparc-oldgcc;
this also means that someone who reads the output of ./config will
notice that it may be advantegeous to upgrade gcc.

25 years agogcc < 2.8 does not support ultrasparc.
Ulf Möller [Tue, 18 May 1999 23:44:38 +0000 (23:44 +0000)]
gcc < 2.8 does not support ultrasparc.

25 years agoChange default PKCS#12 iteration count to 2048, include rsa_oeap_test in the
Dr. Stephen Henson [Tue, 18 May 1999 20:12:54 +0000 (20:12 +0000)]
Change default PKCS#12 iteration count to 2048, include rsa_oeap_test in the
test batch file.

25 years agono-xxx options.
Ulf Möller [Tue, 18 May 1999 16:40:40 +0000 (16:40 +0000)]
no-xxx options.

25 years agoBSD alpha config.
Ulf Möller [Tue, 18 May 1999 16:04:03 +0000 (16:04 +0000)]
BSD alpha config.

Submitted by: Rich Bud <rbud@fortean.com>

25 years agoGet rid of some unnecessary casts and add a necessary one.
Bodo Möller [Tue, 18 May 1999 15:06:20 +0000 (15:06 +0000)]
Get rid of some unnecessary casts and add a necessary one.

25 years agoSpelling error.
Ulf Möller [Tue, 18 May 1999 14:53:39 +0000 (14:53 +0000)]
Spelling error.

25 years agoProtect applications from failing to compile when they
Ralf S. Engelschall [Tue, 18 May 1999 09:19:28 +0000 (09:19 +0000)]
Protect applications from failing to compile when they
try to directly include opensslv.h.

25 years agoDon't forget that it's freeze time now...
Ralf S. Engelschall [Tue, 18 May 1999 08:52:01 +0000 (08:52 +0000)]
Don't forget that it's freeze time now...

25 years agoBC now compiles crypto/des
Ulf Möller [Mon, 17 May 1999 22:33:25 +0000 (22:33 +0000)]
BC now compiles crypto/des

25 years agoMove prototypes to the right place.
Ulf Möller [Mon, 17 May 1999 22:25:31 +0000 (22:25 +0000)]
Move prototypes to the right place.

25 years agoIf we couldn't handle "-showcerts" (which happens with the current
Bodo Möller [Mon, 17 May 1999 20:46:43 +0000 (20:46 +0000)]
If we couldn't handle "-showcerts" (which happens with the current
SSL2 implementation), show at least the server certificate.

25 years agoMention "make depend".
Ulf Möller [Mon, 17 May 1999 20:42:59 +0000 (20:42 +0000)]
Mention "make depend".

25 years agoSubstitute spaces for tab.
Bodo Möller [Mon, 17 May 1999 20:36:08 +0000 (20:36 +0000)]
Substitute spaces for tab.

25 years agoVarious clarifications to extension docs: change the name of literal
Dr. Stephen Henson [Mon, 17 May 1999 20:05:36 +0000 (20:05 +0000)]
Various clarifications to extension docs: change the name of literal
extensions from RAW to DER to avoid confusion with raw extensions.

Update NEWS file.

25 years agoMention Andy's assembler stuff.
Ulf Möller [Mon, 17 May 1999 15:03:43 +0000 (15:03 +0000)]
Mention Andy's assembler stuff.

25 years agomk1mf.pl syntax has been changed.
Ulf Möller [Mon, 17 May 1999 14:38:03 +0000 (14:38 +0000)]
mk1mf.pl syntax has been changed.

25 years agoMake MD5 work on Alpha, and fix a bug.
Bodo Möller [Mon, 17 May 1999 11:21:52 +0000 (11:21 +0000)]
Make MD5 work on Alpha, and fix a bug.
Submitted by: Andy Polyakov

25 years agoAdditional, more descriptive error message for rejection of a session ID
Bodo Möller [Mon, 17 May 1999 11:15:49 +0000 (11:15 +0000)]
Additional, more descriptive error message for rejection of a session ID
because of missing session ID context (so that application programmers
are directly pointed to what they should do differently).

25 years agogcc (in some versions) doesn't like the const_des_cblock typedef.
Bodo Möller [Mon, 17 May 1999 10:54:18 +0000 (10:54 +0000)]
gcc (in some versions) doesn't like the const_des_cblock typedef.
So omit it for now :-(

25 years agoRename "openssl x509" option "-config" to "-extfile", because it
Bodo Möller [Mon, 17 May 1999 08:28:37 +0000 (08:28 +0000)]
Rename "openssl x509" option "-config" to "-extfile", because it
doesn't have a default value like the "-config" options of other
openssl subprograms.

25 years agoClarification.
Bodo Möller [Mon, 17 May 1999 08:17:30 +0000 (08:17 +0000)]
Clarification.

25 years agoOops! Get the pmatch test the right way round.
Dr. Stephen Henson [Sun, 16 May 1999 21:50:26 +0000 (21:50 +0000)]
Oops! Get the pmatch test the right way round.

25 years agoKeep text lines less than 80 characters wide.
Bodo Möller [Sun, 16 May 1999 18:27:46 +0000 (18:27 +0000)]
Keep text lines less than 80 characters wide.

25 years agoFix various less obvious bugs in PKCS#7 handling: such as not zeroing
Dr. Stephen Henson [Sun, 16 May 1999 17:32:32 +0000 (17:32 +0000)]
Fix various less obvious bugs in PKCS#7 handling: such as not zeroing
the secret key before we've encrypted it and using the right NID for RC2-64.
Add various arguments to the experimental programs 'dec' and 'enc' to make
testing less painful.

This stuff has now been tested against Netscape Messenger and it can encrypt
and decrypt S/MIME messages with RC2 (128, 64 and 40 bit) DES and triple DES.

Its still experimental though...

25 years agoTypo.
Bodo Möller [Sun, 16 May 1999 14:20:17 +0000 (14:20 +0000)]
Typo.

25 years agoDES changes.
Bodo Möller [Sun, 16 May 1999 12:29:28 +0000 (12:29 +0000)]
DES changes.

25 years agoChange type of various DES function arguments from des_cblock
Bodo Möller [Sun, 16 May 1999 12:26:16 +0000 (12:26 +0000)]
Change type of various DES function arguments from des_cblock
(meaning pointer to char) to des_cblock * (meaning pointer to
array with 8 char elements), which allows the compiler to
do more typechecking.  (The changed argument types were of type
des_cblock * back in SSLeay, and a lot of ugly casts were
used then to turn them into pointers to elements; but it can be
done without those casts.)

Introduce new type const_des_cblock -- before, the pointers rather
than the elements pointed to were declared const, and for
some reason gcc did not complain about this (but some other
compilers did).

25 years agoAdded a comment pointing out the behaviour of "openssl x509 -conf ...",
Bodo Möller [Sun, 16 May 1999 12:17:20 +0000 (12:17 +0000)]
Added a comment pointing out the behaviour of "openssl x509 -conf ...",
which cost me some time to find out about.

25 years agoMarked probable bug, pointed out by Anonymous.
Bodo Möller [Sun, 16 May 1999 12:01:49 +0000 (12:01 +0000)]
Marked probable bug, pointed out by Anonymous.

25 years agoConvert void *.
Ben Laurie [Sun, 16 May 1999 11:26:46 +0000 (11:26 +0000)]
Convert void *.

25 years agoDon't run "make depend" automatically.
Bodo Möller [Sun, 16 May 1999 10:55:40 +0000 (10:55 +0000)]
Don't run "make depend" automatically.

25 years agoFix some obvious bugs in the PKCS#7 library handling. It didn't try to
Dr. Stephen Henson [Sun, 16 May 1999 00:25:36 +0000 (00:25 +0000)]
Fix some obvious bugs in the PKCS#7 library handling. It didn't try to
find the right RecipientInfo based on the recipient certificate (so would
fail a lot of the time) and fixup cipher structures to correctly (maybe)
modify the AlgorithmIdentifiers.  Largely untested at present... this will be
fixed in due course. Well the stuff was broken to begin with so if its broken
now then you haven't lost anything :-)

25 years agoVarious Win32 fixes. Change args in do_ms.bat to put platform last. Fix
Dr. Stephen Henson [Sat, 15 May 1999 20:33:15 +0000 (20:33 +0000)]
Various Win32 fixes. Change args in do_ms.bat to put platform last. Fix
unsigned/signed cmp error in asn1parse. Change various pem_all.c args to
use pem_password_cb.

25 years agoGet rid of casts.
Ben Laurie [Sat, 15 May 1999 16:03:26 +0000 (16:03 +0000)]
Get rid of casts.

25 years agoAdd actual testing to bntest...
Ben Laurie [Sat, 15 May 1999 15:59:28 +0000 (15:59 +0000)]
Add actual testing to bntest...

25 years agoUse OPENSSL_GLOBAL, OPENSSL_EXTERN instead of GLOBAL and EXTERN.
Bodo Möller [Sat, 15 May 1999 14:38:10 +0000 (14:38 +0000)]
Use OPENSSL_GLOBAL, OPENSSL_EXTERN instead of GLOBAL and EXTERN.

25 years agoOPENSSL_EXTERN
Bodo Möller [Sat, 15 May 1999 14:30:31 +0000 (14:30 +0000)]
OPENSSL_EXTERN

25 years agoOPENSSL_EXTERN, OPENSSL_GLOBAL
Bodo Möller [Sat, 15 May 1999 14:23:29 +0000 (14:23 +0000)]
OPENSSL_EXTERN, OPENSSL_GLOBAL

25 years agoUpdate dependencies.
Bodo Möller [Sat, 15 May 1999 13:38:48 +0000 (13:38 +0000)]
Update dependencies.

25 years agoGet rid of another cast.
Bodo Möller [Sat, 15 May 1999 13:31:53 +0000 (13:31 +0000)]
Get rid of another cast.

25 years agoUse e_os2.h, not e_os.h in exported header file des.h.
Bodo Möller [Sat, 15 May 1999 13:27:13 +0000 (13:27 +0000)]
Use e_os2.h, not e_os.h in exported header file des.h.

25 years agoe_os2.h is used for things that must be visible when external applications
Bodo Möller [Sat, 15 May 1999 13:26:11 +0000 (13:26 +0000)]
e_os2.h is used for things that must be visible when external applications
use (certain parts of) OpenSSL.

25 years agoe_os* corrections.
Bodo Möller [Sat, 15 May 1999 13:24:16 +0000 (13:24 +0000)]
e_os* corrections.

25 years agoComment.
Bodo Möller [Sat, 15 May 1999 12:22:46 +0000 (12:22 +0000)]
Comment.

25 years agoBugfix (set shutdown only when we should).
Bodo Möller [Sat, 15 May 1999 12:22:19 +0000 (12:22 +0000)]
Bugfix (set shutdown only when we should).
Submitted by: Oleg Girko

25 years agoUpdate.
Bodo Möller [Sat, 15 May 1999 12:20:56 +0000 (12:20 +0000)]
Update.

25 years agoGet rid of the cast.
Ben Laurie [Sat, 15 May 1999 11:54:21 +0000 (11:54 +0000)]
Get rid of the cast.

25 years agoReturn 0 for an error, 1 for no error from read_options;
Bodo Möller [Sat, 15 May 1999 11:46:00 +0000 (11:46 +0000)]
Return 0 for an error, 1 for no error from read_options;
that's what the calling code seems to expect.

25 years agoComment.
Bodo Möller [Sat, 15 May 1999 10:40:02 +0000 (10:40 +0000)]
Comment.

25 years agoA comment.
Bodo Möller [Sat, 15 May 1999 00:00:28 +0000 (00:00 +0000)]
A comment.

25 years agoAvoid compiler warnings for x86.
Bodo Möller [Fri, 14 May 1999 21:07:48 +0000 (21:07 +0000)]
Avoid compiler warnings for x86.

25 years agoThe encoding of negative ASN1 INTEGERs and the conversion of BNs to negative
Dr. Stephen Henson [Fri, 14 May 1999 18:21:21 +0000 (18:21 +0000)]
The encoding of negative ASN1 INTEGERs and the conversion of BNs to negative
integers was completely broken. Also added a NEG_PUBKEY_BUG compilation option
to compensate for public keys improperly encoded as negative integers.

25 years agoVarious bugfixes: Uses locking for some more of the stuff that is not
Bodo Möller [Fri, 14 May 1999 12:40:39 +0000 (12:40 +0000)]
Various bugfixes: Uses locking for some more of the stuff that is not
thread-safe (where thread-safe counterparts are not available on all
platforms), and don't memcpy to NULL-pointers
Submitted by: Anonymous
Reviewed by: Bodo Moeller

Also, clean up htons vs. ntohs confusions.

25 years agoFix cryptlib.c.
Bodo Möller [Fri, 14 May 1999 11:57:17 +0000 (11:57 +0000)]
Fix cryptlib.c.
Submitted by:
Reviewed by:
PR:

25 years agoIntroduce and use function typedef pem_password_cb so that we don't call
Bodo Möller [Fri, 14 May 1999 11:52:49 +0000 (11:52 +0000)]
Introduce and use function typedef pem_password_cb so that we don't call
those functions without having a parameter list declaration.
(There are various similar cases left ...)

25 years ago*** empty log message ***
Bodo Möller [Fri, 14 May 1999 11:50:46 +0000 (11:50 +0000)]
*** empty log message ***

25 years agoDelete a wish.
Bodo Möller [Fri, 14 May 1999 11:48:31 +0000 (11:48 +0000)]
Delete a wish.

25 years agoChange cast in function calls to that which is, I think, the right
Bodo Möller [Fri, 14 May 1999 11:47:29 +0000 (11:47 +0000)]
Change cast in function calls to that which is, I think, the right
one for those functions (is it?).

25 years agoAdd release dates to the "CHANGES" file, because that's an obvious
Bodo Möller [Thu, 13 May 1999 21:17:03 +0000 (21:17 +0000)]
Add release dates to the "CHANGES" file, because that's an obvious
place to look for them.

25 years agoMissed one line.
Ulf Möller [Thu, 13 May 1999 17:55:08 +0000 (17:55 +0000)]
Missed one line.

25 years agoDon't use reserved name "_encrypt" for parameters;
Bodo Möller [Thu, 13 May 1999 17:51:35 +0000 (17:51 +0000)]
Don't use reserved name "_encrypt" for parameters;
instead I've picked "enc", because that's what's in the prototypes.
("_encrypt" is reserved only as an external name, but still
using it in an application doesn't look like good style to me --
and it certainly isn't if the point is just avoiding shadowing,
which is apparently why the previous name "encrypt" was changed.)

25 years agoUpdate dependencies.
Ben Laurie [Thu, 13 May 1999 17:33:27 +0000 (17:33 +0000)]
Update dependencies.

25 years agoCut&paste error.
Ulf Möller [Thu, 13 May 1999 17:25:05 +0000 (17:25 +0000)]
Cut&paste error.

25 years agoOops. Close the file.
Ulf Möller [Thu, 13 May 1999 17:24:35 +0000 (17:24 +0000)]
Oops. Close the file.

25 years agoAvoid shadowing, and a bit of constification while I'm at it.
Ben Laurie [Thu, 13 May 1999 17:10:42 +0000 (17:10 +0000)]
Avoid shadowing, and a bit of constification while I'm at it.