]> git.ipfire.org Git - thirdparty/openssl.git/blame_incremental - CHANGES
-N option to diff is not essential, and mentioning it is unnecessarily
[thirdparty/openssl.git] / CHANGES
... / ...
CommitLineData
1
2 OpenSSL CHANGES
3 _______________
4
5 Changes between 0.9.5a and 0.9.6 [xx XXX 2000]
6
7 *) A demo state-machine implementation was sponsored by
8 Nuron (http://www.nuron.com/) and is now available in
9 demos/state_machine.
10 [Ben Laurie]
11
12 *) New options added to the 'dgst' utility for signature
13 generation and verification.
14 [Steve Henson]
15
16 *) Unrecognized PKCS#7 content types are now handled via a
17 catch all ASN1_TYPE structure. This allows unsupported
18 types to be stored as a "blob" and an application can
19 encode and decode it manually.
20 [Steve Henson]
21
22 *) Fix various signed/unsigned issues to make a_strex.c
23 compile under VC++.
24 [Oscar Jacobsson <oscar.jacobsson@celocom.com>]
25
26 *) ASN1 fixes. i2d_ASN1_OBJECT was not returning the correct
27 length if passed a buffer. ASN1_INTEGER_to_BN failed
28 if passed a NULL BN and its argument was negative.
29 [Steve Henson, pointed out by Sven Heiberg <sven@tartu.cyber.ee>]
30
31 *) Modification to PKCS#7 encoding routines to output definite
32 length encoding. Since currently the whole structures are in
33 memory there's not real point in using indefinite length
34 constructed encoding. However if OpenSSL is compiled with
35 the flag PKCS7_INDEFINITE_ENCODING the old form is used.
36 [Steve Henson]
37
38 *) Added BIO_vprintf() and BIO_vsnprintf().
39 [Richard Levitte]
40
41 *) Added more prefixes to parse for in the the strings written
42 through a logging bio, to cover all the levels that are available
43 through syslog. The prefixes are now:
44
45 PANIC, EMERG, EMR => LOG_EMERG
46 ALERT, ALR => LOG_ALERT
47 CRIT, CRI => LOG_CRIT
48 ERROR, ERR => LOG_ERR
49 WARNING, WARN, WAR => LOG_WARNING
50 NOTICE, NOTE, NOT => LOG_NOTICE
51 INFO, INF => LOG_INFO
52 DEBUG, DBG => LOG_DEBUG
53
54 and as before, if none of those prefixes are present at the
55 beginning of the string, LOG_ERR is chosen.
56
57 On Win32, the LOG_* levels are mapped according to this:
58
59 LOG_EMERG, LOG_ALERT, LOG_CRIT, LOG_ERR => EVENTLOG_ERROR_TYPE
60 LOG_WARNING => EVENTLOG_WARNING_TYPE
61 LOG_NOTICE, LOG_INFO, LOG_DEBUG => EVENTLOG_INFORMATION_TYPE
62
63 [Richard Levitte]
64
65 *) Made it possible to reconfigure with just the configuration
66 argument "reconf" or "reconfigure". The command line arguments
67 are stored in Makefile.ssl in the variable CONFIGURE_ARGS,
68 and are retrieved from there when reconfiguring.
69 [Richard Levitte]
70
71 *) MD4 implemented.
72 [Assar Westerlund <assar@sics.se>, Richard Levitte]
73
74 *) Add the arguments -CAfile and -CApath to the pkcs12 utility.
75 [Richard Levitte]
76
77 *) The obj_dat.pl script was messing up the sorting of object
78 names. The reason was that it compared the quoted version
79 of strings as a result "OCSP" > "OCSP Signing" because
80 " > SPACE. Changed script to store unquoted versions of
81 names and add quotes on output. It was also omitting some
82 names from the lookup table if they were given a default
83 value (that is if SN is missing it is given the same
84 value as LN and vice versa), these are now added on the
85 grounds that if an object has a name we should be able to
86 look it up. Finally added warning output when duplicate
87 short or long names are found.
88 [Steve Henson]
89
90 *) Changes needed for Tandem NSK.
91 [Scott Uroff scott@xypro.com]
92
93 *) Fix SSL 2.0 rollback checking: Due to an off-by-one error in
94 RSA_padding_check_SSLv23(), special padding was never detected
95 and thus the SSL 3.0/TLS 1.0 countermeasure against protocol
96 version rollback attacks was not effective.
97
98 In s23_clnt.c, don't use special rollback-attack detection padding
99 (RSA_SSLV23_PADDING) if SSL 2.0 is the only protocol enabled in the
100 client; similarly, in s23_srvr.c, don't do the rollback check if
101 SSL 2.0 is the only protocol enabled in the server.
102 [Bodo Moeller]
103
104 *) Make it possible to get hexdumps of unprintable data with 'openssl
105 asn1parse'. By implication, the functions ASN1_parse_dump() and
106 BIO_dump_indent() are added.
107 [Richard Levitte]
108
109 *) New functions ASN1_STRING_print_ex() and X509_NAME_print_ex()
110 these print out strings and name structures based on various
111 flags including RFC2253 support and proper handling of
112 multibyte characters. Added options to the 'x509' utility
113 to allow the various flags to be set.
114 [Steve Henson]
115
116 *) Various fixes to use ASN1_TIME instead of ASN1_UTCTIME.
117 Also change the functions X509_cmp_current_time() and
118 X509_gmtime_adj() work with an ASN1_TIME structure,
119 this will enable certificates using GeneralizedTime in validity
120 dates to be checked.
121 [Steve Henson]
122
123 *) Make the NEG_PUBKEY_BUG code (which tolerates invalid
124 negative public key encodings) on by default,
125 NO_NEG_PUBKEY_BUG can be set to disable it.
126 [Steve Henson]
127
128 *) New function c2i_ASN1_OBJECT() which acts on ASN1_OBJECT
129 content octets. An i2c_ASN1_OBJECT is unnecessary because
130 the encoding can be trivially obtained from the structure.
131 [Steve Henson]
132
133 *) crypto/err.c locking bugfix: Use write locks (CRYPTO_w_[un]lock),
134 not read locks (CRYPTO_r_[un]lock).
135 [Bodo Moeller]
136
137 *) A first attempt at creating official support for shared
138 libraries through configuration. I've kept it so the
139 default is static libraries only, and the OpenSSL programs
140 are always statically linked for now, but there are
141 preparations for dynamic linking in place.
142 This has been tested on Linux and True64.
143 [Richard Levitte]
144
145 *) Randomness polling function for Win9x, as described in:
146 Peter Gutmann, Software Generation of Practically Strong
147 Random Numbers.
148