test was never triggered due to an off-by-one error.

In s23_clnt.c, don't use special rollback-attack detection padding
(RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
client; similarly, in s23_srvr.c, don't do the rollback check if
SSL 2.0 is the only protocol enabled in the server.

New ASN1_STRING_print_ex() and X509_NAME_print_ex()
functions. These are intended to be replacements
for the ancient ASN1_STRING_print() and X509_NAME_print()
functions.

The new functions support RFC2253 and various pretty
printing options. It is also possible to display
international characters if the terminal properly handles
UTF8 encoding (Linux seems to tolerate this if the
"unicode_start" script is run).

Still needs to be documented, integrated into other
utilities and extensively tested.

therefore have unallocated parts.  Therefore, a check for the 0 case
is needed, resulting with the same thing as when a is negative.

BN_add_word becomes wrongly negative...
This was discovered by Darrel Hankerson <dhankers@cacr.math.uwaterloo.ca>

supporting threads, which means that th assertion is supperbly
dangerous, so make sure it's not compiled under Win32, period.

'openssl asn1parse'.  As a side effect, the functions ASN1_parse_dump
and BIO_dump_indent are added.

didn't recognise NULL to mean 'don't output anything'

size) through the base64 filter, b64_write() messes up it's parameters
in such a way that instead of writing correct base64 output, the first
4 characters of that output is repeated over and over.  This fix
corrects that problem.

it wants to stir the pool using ssleay_rand_add.  This fix provides the
possibility to call ssleay_rand_add inside a locked state by simply telling
it not to do any locking through a static variable.  This isn't the most
elegant way one could do this, but it does retain thread safety during the
stirring process.

In crypto/err/err.c need to initialise p to
NULL in case thread_hash is NULL. Otherwise
p will be uninitialized.

ASN1_TIME fixes.

New function c2i_ASN1_OBJECT().

that of the result pointer to bn_mul was the same as one of the two arguments,
That argument might have it's sign destroyed.  He provided this fix.

state at the same time.

there's support for building under Linux and True64 (using examples
from the programming manuals), including versioning that is currently
the same as OpenSSL versions but should really be a different series.

With this change, it's up to the users to decide if they want shared
libraries as well as the static ones.  This decision now has to be
done at configuration time (well, not really, those who know what they
do can still do it the same way as before).

The OpenSSL programs (openssl and the test programs) are currently
always linked statically, but this may change in the future in a
configurable manner.  The necessary makefile variables to enable this
are in place.

Also note that I have done absolutely nothing about the Windows target
to get something similar.  On the other hand, DLLs are already the
default there, but without versioning, and I've no idea what the
possibilities for such a thing are there...

rand_lcl.h

call the i2c/c2i (they were not using the
content length for the headers).

Fix ASN1 long form tag encoding. This never
worked but it was never tested since it is
only used for tags > 30.

New options to smime program to allow the
PKCS#7 format to be specified and the content
supplied externally.

New ASN1 functions that just deal with
content octets, not tag+length.

could be done automagically, much like the numbering in libeay.num and
ssleay.num.  The solution works as follows:

  - New object identifiers are inserted in objects.txt, following the
    syntax given in objects.README.
  - objects.pl is used to process obj_mac.num and create a new
    obj_mac.h.
  - obj_dat.pl is used to create a new obj_dat.h, using the data in
    obj_mac.h.

This is currently kind of a hack, and the perl code in objects.pl
isn't very elegant, but it works as I intended.  The simplest way to
check that it worked correctly is to look in obj_dat.h and check the
array nid_objs and make sure the objects haven't moved around (this is
important!).  Additions are OK, as well as consistent name changes.

Submitted by: bowe@chip.ma.certco.com

The message to everyone is "Do not hack OpenSSL when stressed"...

the RANDEGD environment variable.

today.  Time to add extra info so the poor users know where to
go with their troubles.

just converting some spaces to tabs from his fix. :-)

comment out in the previous commit

and for small 'a' also a couple of calls to
BN_mod_mul_montgomery(r, r, r, ...).

the 'INSTALL' file, which means that 9 times of 10, the BlowFish
headers won't get installed.  Avoid this in the same way it's done in
crypto/des/Makefile.ssl, where someone apparently has thought of this...

into lexical order. Previously it depended on
the order of files in the directory.

This should now mean that all systems will
agree on the order of safestack.h and will
not change it needlessly and avoid massive
needless commits to safestack.h in future.

It wont however avoid this one :-(

  it cope with OpenBSD which doesn't understand "RTLD_NOW".
* Added the dso_scheme config string entry for OpenBSD-x86 to give it
  DSO support.
* 'make update' that has also absorbed some of Steve's mkstack changes
  for the ASN-related macros.

Fixes for Win32 build.

This is mostly a work around for the old VC++ problem
that it treats func() as func(void).

Various prototypes had been added to 'compare' function
pointers that triggered this. This could be fixed by removing
the prototype, adding function pointer casts to every call or
changing the passed function to use the expected arguments.
I mostly did the latter.

The mkdef.pl script was modified to remove the typesafe
functions which no longer exist.

Oh and some functions called OPENSSL_freeLibrary() were
changed back to FreeLibrary(), wonder how that happened :-)

Handle ASN1_SET_OF and PKCS12_STACK_OF using function
casts in the same way as STACK_OF.