Fix for d2i_ASN1_bytes and stop PKCS#7 routines crashing is signed message

[openssl.git] / CHANGES

 OpenSSL CHANGES
 _______________

 Changes between 0.9.4 and 0.9.5  [xx XXX 1999]

  *) Fix so PKCS7_dataVerify() doesn't crash if no certificates are contained
     in the message. This was handled by allowing
     X509_find_by_issuer_and_serial() to tolerate a NULL passed to it.
     [Steve Henson, reported by Sampo Kellomaki <sampo@mail.neuronio.pt>]

  *) Fix for bug in d2i_ASN1_bytes(): other ASN1 functions add an extra null
     to the end of the strings whereas this didn't. This would cause problems
     if strings read with d2i_ASN1_bytes() were later modified.
     [Steve Henson, reported by Arne Ansper <arne@ats.cyber.ee>]

  *) Fix for base64 decode bug. When a base64 bio reads only one line of
     data and it contains EOF it will end up returning an error. This is
     caused by input 46 bytes long. The cause is due to the way base64
     BIOs find the start of base64 encoded data. They do this by trying a
     trial decode on each line until they find one that works. When they
     do a flag is set and it starts again knowing it can pass all the
     data directly through the decoder. Unfortunately it doesn't reset
     the context it uses. This means that if EOF is reached an attempt
     is made to pass two EOFs through the context and this causes the
     resulting error. This can also cause other problems as well. As is
     usual with these problems it takes *ages* to find and the fix is
     trivial: move one line.
     [Steve Henson, reported by ian@uns.ns.ac.yu (Ivan Nejgebauer) ]

  *) Ugly workaround to get s_client and s_server working under Windows. The
     old code wouldn't work because it needed to select() on sockets and the
     tty (for keypresses and to see if data could be written). Win32 only
     supports select() on sockets so we select() with a 1s timeout on the
     sockets and then see if any characters are waiting to be read, if none
     are present then we retry, we also assume we can always write data to
     the tty. This isn't nice because the code then blocks until we've
     received a complete line of data and it is effectively polling the
     keyboard at 1s intervals: however it's quite a bit better than not
     working at all :-) A dedicated Windows application might handle this
     with an event loop for example.
     [Steve Henson]

  *) Enhance RSA_METHOD structure. Now there are two extra methods, rsa_sign
     and rsa_verify. When the RSA_FLAGS_SIGN_VER option is set these functions
     will be called when RSA_sign() and RSA_verify() are used. This is useful
     if rsa_pub_dec() and rsa_priv_enc() equivalents are not available.
     For this to work properly RSA_public_decrypt() and RSA_private_encrypt()
     should *not* be used: RSA_sign() and RSA_verify() must be used instead.
     This necessitated the support of an extra signature type NID_md5_sha1
     for SSL signatures and modifications to the SSL library to use it instead
     of calling RSA_public_decrypt() and RSA_private_encrypt().
     [Steve Henson]

  *) Add new -verify -CAfile and -CApath options to the crl program, these
     will lookup a CRL issuers certificate and verify the signature in a
     similar way to the verify program. Tidy up the crl program so it
     no longer acesses structures directly. Make the ASN1 CRL parsing a bit
     less strict. It will now permit CRL extensions even if it is not
     a V2 CRL: this will allow it to tolerate some broken CRLs.
     [Steve Henson]

  *) Initialize all non-automatic variables each time one of the openssl
     sub-programs is started (this is necessary as they may be started
     multiple times from the "OpenSSL>" prompt).
     [Lennart Bang, Bodo Moeller]

  *) Preliminary compilation option RSA_NULL which disables RSA crypto without
     removing all other RSA functionality (this is what NO_RSA does). This
     is so (for example) those in the US can disable those operations covered
     by the RSA patent while allowing storage and parsing of RSA keys and RSA
     key generation.
     [Steve Henson]

  *) Non-copying interface to BIO pairs.
     (still largely untested)
     [Bodo Moeller]

  *) New function ANS1_tag2str() to convert an ASN1 tag to a descriptive
     ASCII string. This was handled independently in various places before.
     [Steve Henson]

  *) New functions UTF8_getc() and UTF8_putc() that parse and generate
     UTF8 strings a character at a time.
     [Steve Henson]

  *) Use client_version from client hello to select the protocol
     (s23_srvr.c) and for RSA client key exchange verification
     (s3_srvr.c), as required by the SSL 3.0/TLS 1.0 specifications.
     [Bodo Moeller]

  *) Add various utility functions to handle SPKACs, these were previously
     handled by poking round in the structure internals. Added new function
     NETSCAPE_SPKI_print() to print out SPKAC and a new utility 'spkac' to
     print, verify and generate SPKACs. Based on an original idea from
     Massimiliano Pala <madwolf@comune.modena.it> but extensively modified.
     [Steve Henson]

  *) RIPEMD160 is operational on all platforms and is back in 'make test'.
     [Andy Polyakov]

  *) Allow the config file extension section to be overwritten on the
     command line. Based on an original idea from Massimiliano Pala
     <madwolf@comune.modena.it>. The new option is called -extensions
     and can be applied to ca, req and x509. Also -reqexts to override
     the request extensions in req and -crlexts to override the crl extensions
     in ca.
     [Steve Henson]

  *) Add new feature to the SPKAC handling in ca.  Now you can include
     the same field multiple times by preceding it by "XXXX." for example:
     1.OU="Unit name 1"
     2.OU="Unit name 2"
     this is the same syntax as used in the req config file.
     [Steve Henson]

  *) Allow certificate extensions to be added to certificate requests. These
     are specified in a 'req_extensions' option of the req section of the
     config file. They can be printed out with the -text option to req but
     are otherwise ignored at present.
     [Steve Henson]

  *) Fix a horrible bug in enc_read() in crypto/evp/bio_enc.c: if the first
     data read consists of only the final block it would not decrypted because
     EVP_CipherUpdate() would correctly report zero bytes had been decrypted.
     A misplaced 'break' also meant the decrypted final block might not be
     copied until the next read.
     [Steve Henson]

  *) Initial support for DH_METHOD. Again based on RSA_METHOD. Also added
     a few extra parameters to the DH structure: these will be useful if
     for example we want the value of 'q' or implement X9.42 DH.
     [Steve Henson]

  *) Initial support for DSA_METHOD. This is based on the RSA_METHOD and
     provides hooks that allow the default DSA functions or functions on a
     "per key" basis to be replaced. This allows hardware acceleration and
     hardware key storage to be handled without major modification to the
     library. Also added low level modexp hooks and CRYPTO_EX structure and 
     associated functions.
     [Steve Henson]

  *) Add a new flag to memory BIOs, BIO_FLAG_MEM_RDONLY. This marks the BIO
     as "read only": it can't be written to and the buffer it points to will
     not be freed. Reading from a read only BIO is much more efficient than
     a normal memory BIO. This was added because there are several times when
     an area of memory needs to be read from a BIO. The previous method was
     to create a memory BIO and write the data to it, this results in two
     copies of the data and an O(n^2) reading algorithm. There is a new
     function BIO_new_mem_buf() which creates a read only memory BIO from
     an area of memory. Also modified the PKCS#7 routines to use read only
     memory BIOSs.
     [Steve Henson]

  *) Bugfix: ssl23_get_client_hello did not work properly when called in
     state SSL23_ST_SR_CLNT_HELLO_B, i.e. when the first 7 bytes of
     a SSLv2-compatible client hello for SSLv3 or TLSv1 could be read,
     but a retry condition occured while trying to read the rest.
     [Bodo Moeller]

  *) The PKCS7_ENC_CONTENT_new() function was setting the content type as
     NID_pkcs7_encrypted by default: this was wrong since this should almost
     always be NID_pkcs7_data. Also modified the PKCS7_set_type() to handle
     the encrypted data type: this is a more sensible place to put it and it
     allows the PKCS#12 code to be tidied up that duplicated this
     functionality.
     [Steve Henson]

  *) Changed obj_dat.pl script so it takes its input and output files on
     the command line. This should avoid shell escape redirection problems
     under Win32.
     [Steve Henson]

  *) Initial support for certificate extension requests, these are included
     in things like Xenroll certificate requests. Included functions to allow
     extensions to be obtained and added.
     [Steve Henson]

  *) -crlf option to s_client and s_server for sending newlines as
     CRLF (as required by many protocols).
     [Bodo Moeller]

 Changes between 0.9.3a and 0.9.4  [09 Aug 1999]
  
  *) Install libRSAglue.a when OpenSSL is built with RSAref.
     [Ralf S. Engelschall]

  *) A few more ``#ifndef NO_FP_API / #endif'' pairs for consistency.
     [Andrija Antonijevic <TheAntony2@bigfoot.com>]

  *) Fix -startdate and -enddate (which was missing) arguments to 'ca'
     program.
     [Steve Henson]

  *) New function DSA_dup_DH, which duplicates DSA parameters/keys as
     DH parameters/keys (q is lost during that conversion, but the resulting
     DH parameters contain its length).

     For 1024-bit p, DSA_generate_parameters followed by DSA_dup_DH is
     much faster than DH_generate_parameters (which creates parameters
     where p = 2*q + 1), and also the smaller q makes DH computations
     much more efficient (160-bit exponentiation instead of 1024-bit
     exponentiation); so this provides a convenient way to support DHE
     ciphersuites in SSL/TLS servers (see ssl/ssltest.c).  It is of
     utter importance to use
         SSL_CTX_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
     or
         SSL_set_options(s_ctx, SSL_OP_SINGLE_DH_USE);
     when such DH parameters are used, because otherwise small subgroup
     attacks may become possible!
     [Bodo Moeller]

  *) Avoid memory leak in i2d_DHparams.
     [Bodo Moeller]

  *) Allow the -k option to be used more than once in the enc program:
     this allows the same encrypted message to be read by multiple recipients.
     [Steve Henson]

  *) New function OBJ_obj2txt(buf, buf_len, a, no_name), this converts
     an ASN1_OBJECT to a text string. If the "no_name" parameter is set then
     it will always use the numerical form of the OID, even if it has a short
     or long name.
     [Steve Henson]

  *) Added an extra RSA flag: RSA_FLAG_EXT_PKEY. Previously the rsa_mod_exp
     method only got called if p,q,dmp1,dmq1,iqmp components were present,
     otherwise bn_mod_exp was called. In the case of hardware keys for example
     no private key components need be present and it might store extra data
     in the RSA structure, which cannot be accessed from bn_mod_exp.
     By setting RSA_FLAG_EXT_PKEY rsa_mod_exp will always be called for
     private key operations.
     [Steve Henson]

  *) Added support for SPARC Linux.
     [Andy Polyakov]

  *) pem_password_cb function type incompatibly changed from
          typedef int pem_password_cb(char *buf, int size, int rwflag);
     to
          ....(char *buf, int size, int rwflag, void *userdata);
     so that applications can pass data to their callbacks:
     The PEM[_ASN1]_{read,write}... functions and macros now take an
     additional void * argument, which is just handed through whenever
     the password callback is called.
     [Damien Miller <dmiller@ilogic.com.au>; tiny changes by Bodo Moeller]

     New function SSL_CTX_set_default_passwd_cb_userdata.

     Compatibility note: As many C implementations push function arguments
     onto the stack in reverse order, the new library version is likely to
     interoperate with programs that have been compiled with the old
     pem_password_cb definition (PEM_whatever takes some data that
     happens to be on the stack as its last argument, and the callback
     just ignores this garbage); but there is no guarantee whatsoever that
     this will work.

  *) The -DPLATFORM="\"$(PLATFORM)\"" definition and the similar -DCFLAGS=...
     (both in crypto/Makefile.ssl for use by crypto/cversion.c) caused
     problems not only on Windows, but also on some Unix platforms.
     To avoid problematic command lines, these definitions are now in an
     auto-generated file crypto/buildinf.h (created by crypto/Makefile.ssl
     for standard "make" builds, by util/mk1mf.pl for "mk1mf" builds).
     [Bodo Moeller]

  *) MIPS III/IV assembler module is reimplemented.
     [Andy Polyakov]

  *) More DES library cleanups: remove references to srand/rand and
     delete an unused file.
     [Ulf Möller]

  *) Add support for the the free Netwide assembler (NASM) under Win32,
     since not many people have MASM (ml) and it can be hard to obtain.
     This is currently experimental but it seems to work OK and pass all
     the tests. Check out INSTALL.W32 for info.
     [Steve Henson]

  *) Fix memory leaks in s3_clnt.c: All non-anonymous SSL3/TLS1 connections
     without temporary keys kept an extra copy of the server key,
     and connections with temporary keys did not free everything in case
     of an error.
     [Bodo Moeller]

  *) New function RSA_check_key and new openssl rsa option -check
     for verifying the consistency of RSA keys.
     [Ulf Moeller, Bodo Moeller]

  *) Various changes to make Win32 compile work: 
     1. Casts to avoid "loss of data" warnings in p5_crpt2.c
     2. Change unsigned int to int in b_dump.c to avoid "signed/unsigned
        comparison" warnings.
     3. Add sk_<TYPE>_sort to DEF file generator and do make update.
     [Steve Henson]

  *) Add a debugging option to PKCS#5 v2 key generation function: when
     you #define DEBUG_PKCS5V2 passwords, salts, iteration counts and
     derived keys are printed to stderr.
     [Steve Henson]

  *) Copy the flags in ASN1_STRING_dup().
     [Roman E. Pavlov <pre@mo.msk.ru>]

  *) The x509 application mishandled signing requests containing DSA
     keys when the signing key was also DSA and the parameters didn't match.

     It was supposed to omit the parameters when they matched the signing key:
     the verifying software was then supposed to automatically use the CA's
     parameters if they were absent from the end user certificate.

     Omitting parameters is no longer recommended. The test was also
     the wrong way round! This was probably due to unusual behaviour in
     EVP_cmp_parameters() which returns 1 if the parameters match. 
     This meant that parameters were omitted when they *didn't* match and
     the certificate was useless. Certificates signed with 'ca' didn't have
     this bug.
     [Steve Henson, reported by Doug Erickson <Doug.Erickson@Part.NET>]

  *) Memory leak checking (-DCRYPTO_MDEBUG) had some problems.
     The interface is as follows:
     Applications can use
         CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON) aka MemCheck_start(),
         CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF) aka MemCheck_stop();
     "off" is now the default.
     The library internally uses
         CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE) aka MemCheck_off(),
         CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE) aka MemCheck_on()
     to disable memory-checking temporarily.

     Some inconsistent states that previously were possible (and were
     even the default) are now avoided.

     -DCRYPTO_MDEBUG_TIME is new and additionally stores the current time
     with each memory chunk allocated; this is occasionally more helpful
     than just having a counter.

     -DCRYPTO_MDEBUG_THREAD is also new and adds the thread ID.

     -DCRYPTO_MDEBUG_ALL enables all of the above, plus any future
     extensions.
     [Bodo Moeller]

  *) Introduce "mode" for SSL structures (with defaults in SSL_CTX),
     which largely parallels "options", but is for changing API behaviour,
     whereas "options" are about protocol behaviour.
     Initial "mode" flags are:

     SSL_MODE_ENABLE_PARTIAL_WRITE   Allow SSL_write to report success when
                                     a single record has been written.
     SSL_MODE_ACCEPT_MOVING_WRITE_BUFFER  Don't insist that SSL_write
                                     retries use the same buffer location.
                                     (But all of the contents must be
                                     copied!)
     [Bodo Moeller]

  *) Bugfix: SSL_set_mode ignored its parameter, only SSL_CTX_set_mode
     worked.

  *) Fix problems with no-hmac etc.
     [Ulf Möller, pointed out by Brian Wellington <bwelling@tislabs.com>]

  *) New functions RSA_get_default_method(), RSA_set_method() and
     RSA_get_method(). These allows replacement of RSA_METHODs without having
     to mess around with the internals of an RSA structure.
     [Steve Henson]

  *) Fix memory leaks in DSA_do_sign and DSA_is_prime.
     Also really enable memory leak checks in openssl.c and in some
     test programs.
     [Chad C. Mulligan, Bodo Moeller]

  *) Fix a bug in d2i_ASN1_INTEGER() and i2d_ASN1_INTEGER() which can mess
     up the length of negative integers. This has now been simplified to just
     store the length when it is first determined and use it later, rather
     than trying to keep track of where data is copied and updating it to
     point to the end.
     [Steve Henson, reported by Brien Wheeler
      <bwheeler@authentica-security.com>]

  *) Add a new function PKCS7_signatureVerify. This allows the verification
     of a PKCS#7 signature but with the signing certificate passed to the
     function itself. This contrasts with PKCS7_dataVerify which assumes the
     certificate is present in the PKCS#7 structure. This isn't always the
     case: certificates can be omitted from a PKCS#7 structure and be
     distributed by "out of band" means (such as a certificate database).
     [Steve Henson]

  *) Complete the PEM_* macros with DECLARE_PEM versions to replace the
     function prototypes in pem.h, also change util/mkdef.pl to add the
     necessary function names. 
     [Steve Henson]

  *) mk1mf.pl (used by Windows builds) did not properly read the
     options set by Configure in the top level Makefile, and Configure
     was not even able to write more than one option correctly.
     Fixed, now "no-idea no-rc5 -DCRYPTO_MDEBUG" etc. works as intended.
     [Bodo Moeller]

  *) New functions CONF_load_bio() and CONF_load_fp() to allow a config
     file to be loaded from a BIO or FILE pointer. The BIO version will
     for example allow memory BIOs to contain config info.
     [Steve Henson]

  *) New function "CRYPTO_num_locks" that returns CRYPTO_NUM_LOCKS.
     Whoever hopes to achieve shared-library compatibility across versions
     must use this, not the compile-time macro.
     (Exercise 0.9.4: Which is the minimum library version required by
     such programs?)
     Note: All this applies only to multi-threaded programs, others don't
     need locks.
     [Bodo Moeller]

  *) Add missing case to s3_clnt.c state machine -- one of the new SSL tests
     through a BIO pair triggered the default case, i.e.
     SSLerr(...,SSL_R_UNKNOWN_STATE).
     [Bodo Moeller]

  *) New "BIO pair" concept (crypto/bio/bss_bio.c) so that applications
     can use the SSL library even if none of the specific BIOs is
     appropriate.
     [Bodo Moeller]

  *) Fix a bug in i2d_DSAPublicKey() which meant it returned the wrong value
     for the encoded length.
     [Jeon KyoungHo <khjeon@sds.samsung.co.kr>]

  *) Add initial documentation of the X509V3 functions.
     [Steve Henson]

  *) Add a new pair of functions PEM_write_PKCS8PrivateKey() and 
     PEM_write_bio_PKCS8PrivateKey() that are equivalent to
     PEM_write_PrivateKey() and PEM_write_bio_PrivateKey() but use the more
     secure PKCS#8 private key format with a high iteration count.
     [Steve Henson]

  *) Fix determination of Perl interpreter: A perl or perl5
     _directory_ in $PATH was also accepted as the interpreter.
     [Ralf S. Engelschall]

  *) Fix demos/sign/sign.c: well there wasn't anything strictly speaking
     wrong with it but it was very old and did things like calling
     PEM_ASN1_read() directly and used MD5 for the hash not to mention some
     unusual formatting.
     [Steve Henson]

  *) Fix demos/selfsign.c: it used obsolete and deleted functions, changed
     to use the new extension code.
     [Steve Henson]

  *) Implement the PEM_read/PEM_write functions in crypto/pem/pem_all.c
     with macros. This should make it easier to change their form, add extra
     arguments etc. Fix a few PEM prototypes which didn't have cipher as a
     constant.
     [Steve Henson]

  *) Add to configuration table a new entry that can specify an alternative
     name for unistd.h (for pre-POSIX systems); we need this for NeXTstep,
     according to Mark Crispin <MRC@Panda.COM>.
     [Bodo Moeller]

#if 0
  *) DES CBC did not update the IV. Weird.
     [Ben Laurie]
#else
     des_cbc_encrypt does not update the IV, but des_ncbc_encrypt does.
     Changing the behaviour of the former might break existing programs --
     where IV updating is needed, des_ncbc_encrypt can be used.
#endif

  *) When bntest is run from "make test" it drives bc to check its
     calculations, as well as internally checking them. If an internal check
     fails, it needs to cause bc to give a non-zero result or make test carries
     on without noticing the failure. Fixed.
     [Ben Laurie]

  *) DES library cleanups.
     [Ulf Möller]

  *) Add support for PKCS#5 v2.0 PBE algorithms. This will permit PKCS#8 to be
     used with any cipher unlike PKCS#5 v1.5 which can at most handle 64 bit
     ciphers. NOTE: although the key derivation function has been verified
     against some published test vectors it has not been extensively tested
     yet. Added a -v2 "cipher" option to pkcs8 application to allow the use
     of v2.0.
     [Steve Henson]

  *) Instead of "mkdir -p", which is not fully portable, use new
     Perl script "util/mkdir-p.pl".
     [Bodo Moeller]

  *) Rewrite the way password based encryption (PBE) is handled. It used to
     assume that the ASN1 AlgorithmIdentifier parameter was a PBEParameter
     structure. This was true for the PKCS#5 v1.5 and PKCS#12 PBE algorithms
     but doesn't apply to PKCS#5 v2.0 where it can be something else. Now
     the 'parameter' field of the AlgorithmIdentifier is passed to the
     underlying key generation function so it must do its own ASN1 parsing.
     This has also changed the EVP_PBE_CipherInit() function which now has a
     'parameter' argument instead of literal salt and iteration count values
     and the function EVP_PBE_ALGOR_CipherInit() has been deleted.
     [Steve Henson]

  *) Support for PKCS#5 v1.5 compatible password based encryption algorithms
     and PKCS#8 functionality. New 'pkcs8' application linked to openssl.
     Needed to change the PEM_STRING_EVP_PKEY value which was just "PRIVATE
     KEY" because this clashed with PKCS#8 unencrypted string. Since this
     value was just used as a "magic string" and not used directly its
     value doesn't matter.
     [Steve Henson]

  *) Introduce some semblance of const correctness to BN. Shame C doesn't
     support mutable.
     [Ben Laurie]

  *) "linux-sparc64" configuration (ultrapenguin).
     [Ray Miller <ray.miller@oucs.ox.ac.uk>]
     "linux-sparc" configuration.
     [Christian Forster <fo@hawo.stw.uni-erlangen.de>]

  *) config now generates no-xxx options for missing ciphers.
     [Ulf Möller]

  *) Support the EBCDIC character set (work in progress).
     File ebcdic.c not yet included because it has a different license.
     [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]

  *) Support BS2000/OSD-POSIX.
     [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>]

  *) Make callbacks for key generation use void * instead of char *.
     [Ben Laurie]

  *) Make S/MIME samples compile (not yet tested).
     [Ben Laurie]

  *) Additional typesafe stacks.
     [Ben Laurie]

  *) New configuration variants "bsdi-elf-gcc" (BSD/OS 4.x).
     [Bodo Moeller]


 Changes between 0.9.3 and 0.9.3a  [29 May 1999]

  *) New configuration variant "sco5-gcc".

  *) Updated some demos.
     [Sean O Riordain, Wade Scholine]

  *) Add missing BIO_free at exit of pkcs12 application.
     [Wu Zhigang]

  *) Fix memory leak in conf.c.
     [Steve Henson]

  *) Updates for Win32 to assembler version of MD5.
     [Steve Henson]

  *) Set #! path to perl in apps/der_chop to where we found it
     instead of using a fixed path.
     [Bodo Moeller]

  *) SHA library changes for irix64-mips4-cc.
     [Andy Polyakov]

  *) Improvements for VMS support.
     [Richard Levitte]


 Changes between 0.9.2b and 0.9.3  [24 May 1999]

  *) Bignum library bug fix. IRIX 6 passes "make test" now!
     This also avoids the problems with SC4.2 and unpatched SC5.  
     [Andy Polyakov <appro@fy.chalmers.se>]

  *) New functions sk_num, sk_value and sk_set to replace the previous macros.
     These are required because of the typesafe stack would otherwise break 
     existing code. If old code used a structure member which used to be STACK
     and is now STACK_OF (for example cert in a PKCS7_SIGNED structure) with
     sk_num or sk_value it would produce an error because the num, data members
     are not present in STACK_OF. Now it just produces a warning. sk_set
     replaces the old method of assigning a value to sk_value
     (e.g. sk_value(x, i) = y) which the library used in a few cases. Any code
     that does this will no longer work (and should use sk_set instead) but
     this could be regarded as a "questionable" behaviour anyway.
     [Steve Henson]

  *) Fix most of the other PKCS#7 bugs. The "experimental" code can now
     correctly handle encrypted S/MIME data.
     [Steve Henson]

  *) Change type of various DES function arguments from des_cblock
     (which means, in function argument declarations, pointer to char)
     to des_cblock * (meaning pointer to array with 8 char elements),
     which allows the compiler to do more typechecking; it was like
     that back in SSLeay, but with lots of ugly casts.

     Introduce new type const_des_cblock.
     [Bodo Moeller]

  *) Reorganise the PKCS#7 library and get rid of some of the more obvious
     problems: find RecipientInfo structure that matches recipient certificate
     and initialise the ASN1 structures properly based on passed cipher.
     [Steve Henson]

  *) Belatedly make the BN tests actually check the results.
     [Ben Laurie]

  *) Fix the encoding and decoding of negative ASN1 INTEGERS and conversion
     to and from BNs: it was completely broken. New compilation option
     NEG_PUBKEY_BUG to allow for some broken certificates that encode public
     key elements as negative integers.
     [Steve Henson]

  *) Reorganize and speed up MD5.
     [Andy Polyakov <appro@fy.chalmers.se>]

  *) VMS support.
     [Richard Levitte <richard@levitte.org>]

  *) New option -out to asn1parse to allow the parsed structure to be
     output to a file. This is most useful when combined with the -strparse
     option to examine the output of things like OCTET STRINGS.
     [Steve Henson]

  *) Make SSL library a little more fool-proof by not requiring any longer
     that SSL_set_{accept,connect}_state be called before
     SSL_{accept,connect} may be used (SSL_set_..._state is omitted
     in many applications because usually everything *appeared* to work as
     intended anyway -- now it really works as intended).
     [Bodo Moeller]

  *) Move openssl.cnf out of lib/.
     [Ulf Möller]

  *) Fix various things to let OpenSSL even pass ``egcc -pipe -O2 -Wall
     -Wshadow -Wpointer-arith -Wcast-align -Wmissing-prototypes
     -Wmissing-declarations -Wnested-externs -Winline'' with EGCS 1.1.2+ 
     [Ralf S. Engelschall]

  *) Various fixes to the EVP and PKCS#7 code. It may now be able to
     handle PKCS#7 enveloped data properly.
     [Sebastian Akerman <sak@parallelconsulting.com>, modified by Steve]

  *) Create a duplicate of the SSL_CTX's CERT in SSL_new instead of
     copying pointers.  The cert_st handling is changed by this in
     various ways (and thus what used to be known as ctx->default_cert
     is now called ctx->cert, since we don't resort to s->ctx->[default_]cert
     any longer when s->cert does not give us what we need).
     ssl_cert_instantiate becomes obsolete by this change.
     As soon as we've got the new code right (possibly it already is?),
     we have solved a couple of bugs of the earlier code where s->cert
     was used as if it could not have been shared with other SSL structures.

     Note that using the SSL API in certain dirty ways now will result
     in different behaviour than observed with earlier library versions:
     Changing settings for an SSL_CTX *ctx after having done s = SSL_new(ctx)
     does not influence s as it used to.
     
     In order to clean up things more thoroughly, inside SSL_SESSION
     we don't use CERT any longer, but a new structure SESS_CERT
     that holds per-session data (if available); currently, this is
     the peer's certificate chain and, for clients, the server's certificate
     and temporary key.  CERT holds only those values that can have
     meaningful defaults in an SSL_CTX.
     [Bodo Moeller]

  *) New function X509V3_EXT_i2d() to create an X509_EXTENSION structure
     from the internal representation. Various PKCS#7 fixes: remove some
     evil casts and set the enc_dig_alg field properly based on the signing
     key type.
     [Steve Henson]

  *) Allow PKCS#12 password to be set from the command line or the
     environment. Let 'ca' get its config file name from the environment
     variables "OPENSSL_CONF" or "SSLEAY_CONF" (for consistency with 'req'
     and 'x509').
     [Steve Henson]

  *) Allow certificate policies extension to use an IA5STRING for the
     organization field. This is contrary to the PKIX definition but
     VeriSign uses it and IE5 only recognises this form. Document 'x509'
     extension option.
     [Steve Henson]

  *) Add PEDANTIC compiler flag to allow compilation with gcc -pedantic,
     without disallowing inline assembler and the like for non-pedantic builds.
     [Ben Laurie]

  *) Support Borland C++ builder.
     [Janez Jere <jj@void.si>, modified by Ulf Möller]

  *) Support Mingw32.
     [Ulf Möller]

  *) SHA-1 cleanups and performance enhancements.
     [Andy Polyakov <appro@fy.chalmers.se>]

  *) Sparc v8plus assembler for the bignum library.
     [Andy Polyakov <appro@fy.chalmers.se>]

  *) Accept any -xxx and +xxx compiler options in Configure.
     [Ulf Möller]

  *) Update HPUX configuration.
     [Anonymous]
  
  *) Add missing sk_<type>_unshift() function to safestack.h
     [Ralf S. Engelschall]

  *) New function SSL_CTX_use_certificate_chain_file that sets the
     "extra_cert"s in addition to the certificate.  (This makes sense
     only for "PEM" format files, as chains as a whole are not
     DER-encoded.)
     [Bodo Moeller]

  *) Support verify_depth from the SSL API.
     x509_vfy.c had what can be considered an off-by-one-error:
     Its depth (which was not part of the external interface)
     was actually counting the number of certificates in a chain;
     now it really counts the depth.
     [Bodo Moeller]

  *) Bugfix in crypto/x509/x509_cmp.c: The SSLerr macro was used
     instead of X509err, which often resulted in confusing error
     messages since the error codes are not globally unique
     (e.g. an alleged error in ssl3_accept when a certificate
     didn't match the private key).

  *) New function SSL_CTX_set_session_id_context that allows to set a default
     value (so that you don't need SSL_set_session_id_context for each
     connection using the SSL_CTX).
     [Bodo Moeller]

  *) OAEP decoding bug fix.
     [Ulf Möller]

  *) Support INSTALL_PREFIX for package builders, as proposed by
     David Harris.
     [Bodo Moeller]

  *) New Configure options "threads" and "no-threads".  For systems
     where the proper compiler options are known (currently Solaris
     and Linux), "threads" is the default.
     [Bodo Moeller]

  *) New script util/mklink.pl as a faster substitute for util/mklink.sh.
     [Bodo Moeller]

  *) Install various scripts to $(OPENSSLDIR)/misc, not to
     $(INSTALLTOP)/bin -- they shouldn't clutter directories
     such as /usr/local/bin.
     [Bodo Moeller]

  *) "make linux-shared" to build shared libraries.
     [Niels Poppe <niels@netbox.org>]

  *) New Configure option no-<cipher> (rsa, idea, rc5, ...).
     [Ulf Möller]

  *) Add the PKCS#12 API documentation to openssl.txt. Preliminary support for
     extension adding in x509 utility.
     [Steve Henson]

  *) Remove NOPROTO sections and error code comments.
     [Ulf Möller]

  *) Partial rewrite of the DEF file generator to now parse the ANSI
     prototypes.
     [Steve Henson]

  *) New Configure options --prefix=DIR and --openssldir=DIR.
     [Ulf Möller]

  *) Complete rewrite of the error code script(s). It is all now handled
     by one script at the top level which handles error code gathering,
     header rewriting and C source file generation. It should be much better
     than the old method: it now uses a modified version of Ulf's parser to
     read the ANSI prototypes in all header files (thus the old K&R definitions
     aren't needed for error creation any more) and do a better job of
     translating function codes into names. The old 'ASN1 error code imbedded
     in a comment' is no longer necessary and it doesn't use .err files which
     have now been deleted. Also the error code call doesn't have to appear all
     on one line (which resulted in some large lines...).
     [Steve Henson]

  *) Change #include filenames from <foo.h> to <openssl/foo.h>.
     [Bodo Moeller]

  *) Change behaviour of ssl2_read when facing length-0 packets: Don't return
     0 (which usually indicates a closed connection), but continue reading.
     [Bodo Moeller]

  *) Fix some race conditions.
     [Bodo Moeller]

  *) Add support for CRL distribution points extension. Add Certificate
     Policies and CRL distribution points documentation.
     [Steve Henson]

  *) Move the autogenerated header file parts to crypto/opensslconf.h.
     [Ulf Möller]

  *) Fix new 56-bit DES export ciphersuites: they were using 7 bytes instead of
     8 of keying material. Merlin has also confirmed interop with this fix
     between OpenSSL and Baltimore C/SSL 2.0 and J/SSL 2.0.
     [Merlin Hughes <merlin@baltimore.ie>]

  *) Fix lots of warnings.
     [Richard Levitte <levitte@stacken.kth.se>]
 
  *) In add_cert_dir() in crypto/x509/by_dir.c, break out of the loop if
     the directory spec didn't end with a LIST_SEPARATOR_CHAR.
     [Richard Levitte <levitte@stacken.kth.se>]
 
  *) Fix problems with sizeof(long) == 8.
     [Andy Polyakov <appro@fy.chalmers.se>]

  *) Change functions to ANSI C.
     [Ulf Möller]

  *) Fix typos in error codes.
     [Martin Kraemer <Martin.Kraemer@MchP.Siemens.De>, Ulf Möller]

  *) Remove defunct assembler files from Configure.
     [Ulf Möller]

  *) SPARC v8 assembler BIGNUM implementation.
     [Andy Polyakov <appro@fy.chalmers.se>]

  *) Support for Certificate Policies extension: both print and set.
     Various additions to support the r2i method this uses.
     [Steve Henson]

  *) A lot of constification, and fix a bug in X509_NAME_oneline() that could
     return a const string when you are expecting an allocated buffer.
     [Ben Laurie]

  *) Add support for ASN1 types UTF8String and VISIBLESTRING, also the CHOICE
     types DirectoryString and DisplayText.
     [Steve Henson]

  *) Add code to allow r2i extensions to access the configuration database,
     add an LHASH database driver and add several ctx helper functions.
     [Steve Henson]

  *) Fix an evil bug in bn_expand2() which caused various BN functions to
     fail when they extended the size of a BIGNUM.
     [Steve Henson]

  *) Various utility functions to handle SXNet extension. Modify mkdef.pl to
     support typesafe stack.
     [Steve Henson]

  *) Fix typo in SSL_[gs]et_options().
     [Nils Frostberg <nils@medcom.se>]

  *) Delete various functions and files that belonged to the (now obsolete)
     old X509V3 handling code.
     [Steve Henson]

  *) New Configure option "rsaref".
     [Ulf Möller]

  *) Don't auto-generate pem.h.
     [Bodo Moeller]

  *) Introduce type-safe ASN.1 SETs.
     [Ben Laurie]

  *) Convert various additional casted stacks to type-safe STACK_OF() variants.
     [Ben Laurie, Ralf S. Engelschall, Steve Henson]

  *) Introduce type-safe STACKs. This will almost certainly break lots of code
     that links with OpenSSL (well at least cause lots of warnings), but fear
     not: the conversion is trivial, and it eliminates loads of evil casts. A
     few STACKed things have been converted already. Feel free to convert more.
     In the fullness of time, I'll do away with the STACK type altogether.
     [Ben Laurie]

  *) Add `openssl ca -revoke <certfile>' facility which revokes a certificate
     specified in <certfile> by updating the entry in the index.txt file.
     This way one no longer has to edit the index.txt file manually for
     revoking a certificate. The -revoke option does the gory details now.
     [Massimiliano Pala <madwolf@openca.org>, Ralf S. Engelschall]

  *) Fix `openssl crl -noout -text' combination where `-noout' killed the
     `-text' option at all and this way the `-noout -text' combination was
     inconsistent in `openssl crl' with the friends in `openssl x509|rsa|dsa'.
     [Ralf S. Engelschall]

  *) Make sure a corresponding plain text error message exists for the
     X509_V_ERR_CERT_REVOKED/23 error number which can occur when a
     verify callback function determined that a certificate was revoked.
     [Ralf S. Engelschall]

  *) Bugfix: In test/testenc, don't test "openssl <cipher>" for
     ciphers that were excluded, e.g. by -DNO_IDEA.  Also, test
     all available cipers including rc5, which was forgotten until now.
     In order to let the testing shell script know which algorithms
     are available, a new (up to now undocumented) command
     "openssl list-cipher-commands" is used.
     [Bodo Moeller]

  *) Bugfix: s_client occasionally would sleep in select() when
     it should have checked SSL_pending() first.
     [Bodo Moeller]

  *) New functions DSA_do_sign and DSA_do_verify to provide access to
     the raw DSA values prior to ASN.1 encoding.
     [Ulf Möller]

  *) Tweaks to Configure
     [Niels Poppe <niels@netbox.org>]

  *) Add support for PKCS#5 v2.0 ASN1 PBES2 structures. No other support,
     yet...
     [Steve Henson]

  *) New variables $(RANLIB) and $(PERL) in the Makefiles.
     [Ulf Möller]

  *) New config option to avoid instructions that are illegal on the 80386.
     The default code is faster, but requires at least a 486.
     [Ulf Möller]
  
  *) Got rid of old SSL2_CLIENT_VERSION (inconsistently used) and
     SSL2_SERVER_VERSION (not used at all) macros, which are now the
     same as SSL2_VERSION anyway.
     [Bodo Moeller]

  *) New "-showcerts" option for s_client.
     [Bodo Moeller]

  *) Still more PKCS#12 integration. Add pkcs12 application to openssl
     application. Various cleanups and fixes.
     [Steve Henson]

  *) More PKCS#12 integration. Add new pkcs12 directory with Makefile.ssl and
     modify error routines to work internally. Add error codes and PBE init
     to library startup routines.
     [Steve Henson]

  *) Further PKCS#12 integration. Added password based encryption, PKCS#8 and
     packing functions to asn1 and evp. Changed function names and error
     codes along the way.
     [Steve Henson]

  *) PKCS12 integration: and so it begins... First of several patches to
     slowly integrate PKCS#12 functionality into OpenSSL. Add PKCS#12
     objects to objects.h
     [Steve Henson]

  *) Add a new 'indent' option to some X509V3 extension code. Initial ASN1
     and display support for Thawte strong extranet extension.
     [Steve Henson]

  *) Add LinuxPPC support.
     [Jeff Dubrule <igor@pobox.org>]

  *) Get rid of redundant BN file bn_mulw.c, and rename bn_div64 to
     bn_div_words in alpha.s.
     [Hannes Reinecke <H.Reinecke@hw.ac.uk> and Ben Laurie]

  *) Make sure the RSA OAEP test is skipped under -DRSAref because
     OAEP isn't supported when OpenSSL is built with RSAref.
     [Ulf Moeller <ulf@fitug.de>]

  *) Move definitions of IS_SET/IS_SEQUENCE inside crypto/asn1/asn1.h 
     so they no longer are missing under -DNOPROTO. 
     [Soren S. Jorvang <soren@t.dk>]


 Changes between 0.9.1c and 0.9.2b  [22 Mar 1999]

  *) Make SSL_get_peer_cert_chain() work in servers. Unfortunately, it still
     doesn't work when the session is reused. Coming soon!
     [Ben Laurie]

  *) Fix a security hole, that allows sessions to be reused in the wrong
     context thus bypassing client cert protection! All software that uses
     client certs and session caches in multiple contexts NEEDS PATCHING to
     allow session reuse! A fuller solution is in the works.
     [Ben Laurie, problem pointed out by Holger Reif, Bodo Moeller (and ???)]

  *) Some more source tree cleanups (removed obsolete files
     crypto/bf/asm/bf586.pl, test/test.txt and crypto/sha/asm/f.s; changed
     permission on "config" script to be executable) and a fix for the INSTALL
     document.
     [Ulf Moeller <ulf@fitug.de>]

  *) Remove some legacy and erroneous uses of malloc, free instead of
     Malloc, Free.
     [Lennart Bang <lob@netstream.se>, with minor changes by Steve]

  *) Make rsa_oaep_test return non-zero on error.
     [Ulf Moeller <ulf@fitug.de>]

  *) Add support for native Solaris shared libraries. Configure
     solaris-sparc-sc4-pic, make, then run shlib/solaris-sc4.sh. It'd be nice
     if someone would make that last step automatic.
     [Matthias Loepfe <Matthias.Loepfe@AdNovum.CH>]

  *) ctx_size was not built with the right compiler during "make links". Fixed.
     [Ben Laurie]

  *) Change the meaning of 'ALL' in the cipher list. It now means "everything
     except NULL ciphers". This means the default cipher list will no longer
     enable NULL ciphers. They need to be specifically enabled e.g. with
     the string "DEFAULT:eNULL".
     [Steve Henson]

  *) Fix to RSA private encryption routines: if p < q then it would
     occasionally produce an invalid result. This will only happen with
     externally generated keys because OpenSSL (and SSLeay) ensure p > q.
     [Steve Henson]

  *) Be less restrictive and allow also `perl util/perlpath.pl
     /path/to/bin/perl' in addition to `perl util/perlpath.pl /path/to/bin',
     because this way one can also use an interpreter named `perl5' (which is
     usually the name of Perl 5.xxx on platforms where an Perl 4.x is still
     installed as `perl').
     [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]

  *) Let util/clean-depend.pl work also with older Perl 5.00x versions.
     [Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]

  *) Fix Makefile.org so CC,CFLAG etc are passed to 'make links' add
     advapi32.lib to Win32 build and change the pem test comparision
     to fc.exe (thanks to Ulrich Kroener <kroneru@yahoo.com> for the
     suggestion). Fix misplaced ASNI prototypes and declarations in evp.h
     and crypto/des/ede_cbcm_enc.c.
     [Steve Henson]

  *) DES quad checksum was broken on big-endian architectures. Fixed.
     [Ben Laurie]

  *) Comment out two functions in bio.h that aren't implemented. Fix up the
     Win32 test batch file so it (might) work again. The Win32 test batch file
     is horrible: I feel ill....
     [Steve Henson]

  *) Move various #ifdefs around so NO_SYSLOG, NO_DIRENT etc are now selected
     in e_os.h. Audit of header files to check ANSI and non ANSI
     sections: 10 functions were absent from non ANSI section and not exported
     from Windows DLLs. Fixed up libeay.num for new functions.
     [Steve Henson]

  *) Make `openssl version' output lines consistent.
     [Ralf S. Engelschall]

  *) Fix Win32 symbol export lists for BIO functions: Added
     BIO_get_ex_new_index, BIO_get_ex_num, BIO_get_ex_data and BIO_set_ex_data
     to ms/libeay{16,32}.def.
     [Ralf S. Engelschall]

  *) Second round of fixing the OpenSSL perl/ stuff. It now at least compiled
     fine under Unix and passes some trivial tests I've now added. But the
     whole stuff is horribly incomplete, so a README.1ST with a disclaimer was
     added to make sure no one expects that this stuff really works in the
     OpenSSL 0.9.2 release.  Additionally I've started to clean the XS sources
     up and fixed a few little bugs and inconsistencies in OpenSSL.{pm,xs} and
     openssl_bio.xs.
     [Ralf S. Engelschall]

  *) Fix the generation of two part addresses in perl.
     [Kenji Miyake <kenji@miyake.org>, integrated by Ben Laurie]

  *) Add config entry for Linux on MIPS.
     [John Tobey <jtobey@channel1.com>]

  *) Make links whenever Configure is run, unless we are on Windoze.
     [Ben Laurie]

  *) Permit extensions to be added to CRLs using crl_section in openssl.cnf.
     Currently only issuerAltName and AuthorityKeyIdentifier make any sense
     in CRLs.
     [Steve Henson]

  *) Add a useful kludge to allow package maintainers to specify compiler and
     other platforms details on the command line without having to patch the
     Configure script everytime: One now can use ``perl Configure
     <id>:<details>'', i.e. platform ids are allowed to have details appended
     to them (seperated by colons). This is treated as there would be a static
     pre-configured entry in Configure's %table under key <id> with value
     <details> and ``perl Configure <id>'' is called.  So, when you want to
     perform a quick test-compile under FreeBSD 3.1 with pgcc and without
     assembler stuff you can use ``perl Configure "FreeBSD-elf:pgcc:-O6:::"''
     now, which overrides the FreeBSD-elf entry on-the-fly.
     [Ralf S. Engelschall]

  *) Disable new TLS1 ciphersuites by default: they aren't official yet.
     [Ben Laurie]

  *) Allow DSO flags like -fpic, -fPIC, -KPIC etc. to be specified
     on the `perl Configure ...' command line. This way one can compile
     OpenSSL libraries with Position Independent Code (PIC) which is needed
     for linking it into DSOs.
     [Ralf S. Engelschall]

  *) Remarkably, export ciphers were totally broken and no-one had noticed!
     Fixed.
     [Ben Laurie]

  *) Cleaned up the LICENSE document: The official contact for any license
     questions now is the OpenSSL core team under openssl-core@openssl.org.
     And add a paragraph about the dual-license situation to make sure people
     recognize that _BOTH_ the OpenSSL license _AND_ the SSLeay license apply
     to the OpenSSL toolkit.
     [Ralf S. Engelschall]

  *) General source tree makefile cleanups: Made `making xxx in yyy...'
     display consistent in the source tree and replaced `/bin/rm' by `rm'.
     Additonally cleaned up the `make links' target: Remove unnecessary
     semicolons, subsequent redundant removes, inline point.sh into mklink.sh
     to speed processing and no longer clutter the display with confusing
     stuff. Instead only the actually done links are displayed.
     [Ralf S. Engelschall]

  *) Permit null encryption ciphersuites, used for authentication only. It used
     to be necessary to set the preprocessor define SSL_ALLOW_ENULL to do this.
     It is now necessary to set SSL_FORBID_ENULL to prevent the use of null
     encryption.
     [Ben Laurie]

  *) Add a bunch of fixes to the PKCS#7 stuff. It used to sometimes reorder
     signed attributes when verifying signatures (this would break them), 
     the detached data encoding was wrong and public keys obtained using
     X509_get_pubkey() weren't freed.
     [Steve Henson]

  *) Add text documentation for the BUFFER functions. Also added a work around
     to a Win95 console bug. This was triggered by the password read stuff: the
     last character typed gets carried over to the next fread(). If you were 
     generating a new cert request using 'req' for example then the last
     character of the passphrase would be CR which would then enter the first
     field as blank.
     [Steve Henson]

  *) Added the new `Includes OpenSSL Cryptography Software' button as
     doc/openssl_button.{gif,html} which is similar in style to the old SSLeay
     button and can be used by applications based on OpenSSL to show the
     relationship to the OpenSSL project.  
     [Ralf S. Engelschall]

  *) Remove confusing variables in function signatures in files
     ssl/ssl_lib.c and ssl/ssl.h.
     [Lennart Bong <lob@kulthea.stacken.kth.se>]

  *) Don't install bss_file.c under PREFIX/include/
     [Lennart Bong <lob@kulthea.stacken.kth.se>]

  *) Get the Win32 compile working again. Modify mkdef.pl so it can handle
     functions that return function pointers and has support for NT specific
     stuff. Fix mk1mf.pl and VC-32.pl to support NT differences also. Various
     #ifdef WIN32 and WINNTs sprinkled about the place and some changes from
     unsigned to signed types: this was killing the Win32 compile.
     [Steve Henson]

  *) Add new certificate file to stack functions,
     SSL_add_dir_cert_subjects_to_stack() and
     SSL_add_file_cert_subjects_to_stack().  These largely supplant
     SSL_load_client_CA_file(), and can be used to add multiple certs easily
     to a stack (usually this is then handed to SSL_CTX_set_client_CA_list()).
     This means that Apache-SSL and similar packages don't have to mess around
     to add as many CAs as they want to the preferred list.
     [Ben Laurie]

  *) Experiment with doxygen documentation. Currently only partially applied to
     ssl/ssl_lib.c.
     See http://www.stack.nl/~dimitri/doxygen/index.html, and run doxygen with
     openssl.doxy as the configuration file.
     [Ben Laurie]
  
  *) Get rid of remaining C++-style comments which strict C compilers hate.
     [Ralf S. Engelschall, pointed out by Carlos Amengual]

  *) Changed BN_RECURSION in bn_mont.c to BN_RECURSION_MONT so it is not
     compiled in by default: it has problems with large keys.
     [Steve Henson]

  *) Add a bunch of SSL_xxx() functions for configuring the temporary RSA and
     DH private keys and/or callback functions which directly correspond to
     their SSL_CTX_xxx() counterparts but work on a per-connection basis. This
     is needed for applications which have to configure certificates on a
     per-connection basis (e.g. Apache+mod_ssl) instead of a per-context basis
     (e.g. s_server). 
        For the RSA certificate situation is makes no difference, but
     for the DSA certificate situation this fixes the "no shared cipher"
     problem where the OpenSSL cipher selection procedure failed because the
     temporary keys were not overtaken from the context and the API provided
     no way to reconfigure them. 
        The new functions now let applications reconfigure the stuff and they
     are in detail: SSL_need_tmp_RSA, SSL_set_tmp_rsa, SSL_set_tmp_dh,
     SSL_set_tmp_rsa_callback and SSL_set_tmp_dh_callback.  Additionally a new
     non-public-API function ssl_cert_instantiate() is used as a helper
     function and also to reduce code redundancy inside ssl_rsa.c.
     [Ralf S. Engelschall]

  *) Move s_server -dcert and -dkey options out of the undocumented feature
     area because they are useful for the DSA situation and should be
     recognized by the users.
     [Ralf S. Engelschall]

  *) Fix the cipher decision scheme for export ciphers: the export bits are
     *not* within SSL_MKEY_MASK or SSL_AUTH_MASK, they are within
     SSL_EXP_MASK.  So, the original variable has to be used instead of the
     already masked variable.
     [Richard Levitte <levitte@stacken.kth.se>]

  *) Fix 'port' variable from `int' to `unsigned int' in crypto/bio/b_sock.c
     [Richard Levitte <levitte@stacken.kth.se>]

  *) Change type of another md_len variable in pk7_doit.c:PKCS7_dataFinal()
     from `int' to `unsigned int' because it's a length and initialized by
     EVP_DigestFinal() which expects an `unsigned int *'.
     [Richard Levitte <levitte@stacken.kth.se>]

  *) Don't hard-code path to Perl interpreter on shebang line of Configure
     script. Instead use the usual Shell->Perl transition trick.
     [Ralf S. Engelschall]

  *) Make `openssl x509 -noout -modulus' functional also for DSA certificates
     (in addition to RSA certificates) to match the behaviour of `openssl dsa
     -noout -modulus' as it's already the case for `openssl rsa -noout
     -modulus'.  For RSA the -modulus is the real "modulus" while for DSA
     currently the public key is printed (a decision which was already done by
     `openssl dsa -modulus' in the past) which serves a similar purpose.
     Additionally the NO_RSA no longer completely removes the whole -modulus
     option; it now only avoids using the RSA stuff. Same applies to NO_DSA
     now, too.
     [Ralf S.  Engelschall]

  *) Add Arne Ansper's reliable BIO - this is an encrypted, block-digested
     BIO. See the source (crypto/evp/bio_ok.c) for more info.
     [Arne Ansper <arne@ats.cyber.ee>]

  *) Dump the old yucky req code that tried (and failed) to allow raw OIDs
     to be added. Now both 'req' and 'ca' can use new objects defined in the
     config file.
     [Steve Henson]

  *) Add cool BIO that does syslog (or event log on NT).
     [Arne Ansper <arne@ats.cyber.ee>, integrated by Ben Laurie]

  *) Add support for new TLS ciphersuites, TLS_RSA_EXPORT56_WITH_RC4_56_MD5,
     TLS_RSA_EXPORT56_WITH_RC2_CBC_56_MD5 and
     TLS_RSA_EXPORT56_WITH_DES_CBC_SHA, as specified in "56-bit Export Cipher
     Suites For TLS", draft-ietf-tls-56-bit-ciphersuites-00.txt.
     [Ben Laurie]

  *) Add preliminary config info for new extension code.
     [Steve Henson]

  *) Make RSA_NO_PADDING really use no padding.
     [Ulf Moeller <ulf@fitug.de>]

  *) Generate errors when private/public key check is done.
     [Ben Laurie]

  *) Overhaul for 'crl' utility. New function X509_CRL_print. Partial support
     for some CRL extensions and new objects added.
     [Steve Henson]

  *) Really fix the ASN1 IMPLICIT bug this time... Partial support for private
     key usage extension and fuller support for authority key id.
     [Steve Henson]

  *) Add OAEP encryption for the OpenSSL crypto library. OAEP is the improved
     padding method for RSA, which is recommended for new applications in PKCS
     #1 v2.0 (RFC 2437, October 1998).
     OAEP (Optimal Asymmetric Encryption Padding) has better theoretical
     foundations than the ad-hoc padding used in PKCS #1 v1.5. It is secure
     against Bleichbacher's attack on RSA.
     [Ulf Moeller <ulf@fitug.de>, reformatted, corrected and integrated by
      Ben Laurie]

  *) Updates to the new SSL compression code
     [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]

  *) Fix so that the version number in the master secret, when passed
     via RSA, checks that if TLS was proposed, but we roll back to SSLv3
     (because the server will not accept higher), that the version number
     is 0x03,0x01, not 0x03,0x00
     [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]

  *) Run extensive memory leak checks on SSL apps. Fixed *lots* of memory
     leaks in ssl/ relating to new X509_get_pubkey() behaviour. Also fixes
     in apps/ and an unrelated leak in crypto/dsa/dsa_vrf.c
     [Steve Henson]

  *) Support for RAW extensions where an arbitrary extension can be
     created by including its DER encoding. See apps/openssl.cnf for
     an example.
     [Steve Henson]

  *) Make sure latest Perl versions don't interpret some generated C array
     code as Perl array code in the crypto/err/err_genc.pl script.
     [Lars Weber <3weber@informatik.uni-hamburg.de>]

  *) Modify ms/do_ms.bat to not generate assembly language makefiles since
     not many people have the assembler. Various Win32 compilation fixes and
     update to the INSTALL.W32 file with (hopefully) more accurate Win32
     build instructions.
     [Steve Henson]

  *) Modify configure script 'Configure' to automatically create crypto/date.h
     file under Win32 and also build pem.h from pem.org. New script
     util/mkfiles.pl to create the MINFO file on environments that can't do a
     'make files': perl util/mkfiles.pl >MINFO should work.
     [Steve Henson]

  *) Major rework of DES function declarations, in the pursuit of correctness
     and purity. As a result, many evil casts evaporated, and some weirdness,
     too. You may find this causes warnings in your code. Zapping your evil
     casts will probably fix them. Mostly.
     [Ben Laurie]

  *) Fix for a typo in asn1.h. Bug fix to object creation script
     obj_dat.pl. It considered a zero in an object definition to mean
     "end of object": none of the objects in objects.h have any zeros
     so it wasn't spotted.
     [Steve Henson, reported by Erwann ABALEA <eabalea@certplus.com>]

  *) Add support for Triple DES Cipher Block Chaining with Output Feedback
     Masking (CBCM). In the absence of test vectors, the best I have been able
     to do is check that the decrypt undoes the encrypt, so far. Send me test
     vectors if you have them.
     [Ben Laurie]

  *) Correct calculation of key length for export ciphers (too much space was
     allocated for null ciphers). This has not been tested!
     [Ben Laurie]

  *) Modifications to the mkdef.pl for Win32 DEF file creation. The usage
     message is now correct (it understands "crypto" and "ssl" on its
     command line). There is also now an "update" option. This will update
     the util/ssleay.num and util/libeay.num files with any new functions.
     If you do a: 
     perl util/mkdef.pl crypto ssl update
     it will update them.
     [Steve Henson]

  *) Overhauled the Perl interface (perl/*):
     - ported BN stuff to OpenSSL's different BN library
     - made the perl/ source tree CVS-aware
     - renamed the package from SSLeay to OpenSSL (the files still contain
       their history because I've copied them in the repository)
     - removed obsolete files (the test scripts will be replaced
       by better Test::Harness variants in the future)
     [Ralf S. Engelschall]

  *) First cut for a very conservative source tree cleanup:
     1. merge various obsolete readme texts into doc/ssleay.txt
     where we collect the old documents and readme texts.
     2. remove the first part of files where I'm already sure that we no
     longer need them because of three reasons: either they are just temporary
     files which were left by Eric or they are preserved original files where
     I've verified that the diff is also available in the CVS via "cvs diff
     -rSSLeay_0_8_1b" or they were renamed (as it was definitely the case for
     the crypto/md/ stuff).
     [Ralf S. Engelschall]

  *) More extension code. Incomplete support for subject and issuer alt
     name, issuer and authority key id. Change the i2v function parameters
     and add an extra 'crl' parameter in the X509V3_CTX structure: guess
     what that's for :-) Fix to ASN1 macro which messed up
     IMPLICIT tag and add f_enum.c which adds a2i, i2a for ENUMERATED.
     [Steve Henson]

  *) Preliminary support for ENUMERATED type. This is largely copied from the
     INTEGER code.
     [Steve Henson]

  *) Add new function, EVP_MD_CTX_copy() to replace frequent use of memcpy.
     [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]

  *) Make sure `make rehash' target really finds the `openssl' program.
     [Ralf S. Engelschall, Matthias Loepfe <Matthias.Loepfe@adnovum.ch>]

  *) Squeeze another 7% of speed out of MD5 assembler, at least on a P2. I'd
     like to hear about it if this slows down other processors.
     [Ben Laurie]

  *) Add CygWin32 platform information to Configure script.
     [Alan Batie <batie@aahz.jf.intel.com>]

  *) Fixed ms/32all.bat script: `no_asm' -> `no-asm'
     [Rainer W. Gerling <gerling@mpg-gv.mpg.de>]
  
  *) New program nseq to manipulate netscape certificate sequences
     [Steve Henson]

  *) Modify crl2pkcs7 so it supports multiple -certfile arguments. Fix a
     few typos.
     [Steve Henson]

  *) Fixes to BN code.  Previously the default was to define BN_RECURSION
     but the BN code had some problems that would cause failures when
     doing certificate verification and some other functions.
     [Eric A. Young, (from changes to C2Net SSLeay, integrated by Mark Cox)]

  *) Add ASN1 and PEM code to support netscape certificate sequences.
     [Steve Henson]

  *) Add ASN1 and PEM code to support netscape certificate sequences.
     [Steve Henson]

  *) Add several PKIX and private extended key usage OIDs.
     [Steve Henson]

  *) Modify the 'ca' program to handle the new extension code. Modify
     openssl.cnf for new extension format, add comments.
     [Steve Henson]

  *) More X509 V3 changes. Fix typo in v3_bitstr.c. Add support to 'req'
     and add a sample to openssl.cnf so req -x509 now adds appropriate
     CA extensions.
     [Steve Henson]

  *) Continued X509 V3 changes. Add to other makefiles, integrate with the
     error code, add initial support to X509_print() and x509 application.
     [Steve Henson]

  *) Takes a deep breath and start addding X509 V3 extension support code. Add
     files in crypto/x509v3. Move original stuff to crypto/x509v3/old. All this
     stuff is currently isolated and isn't even compiled yet.
     [Steve Henson]

  *) Continuing patches for GeneralizedTime. Fix up certificate and CRL
     ASN1 to use ASN1_TIME and modify print routines to use ASN1_TIME_print.
     Removed the versions check from X509 routines when loading extensions:
     this allows certain broken certificates that don't set the version
     properly to be processed.
     [Steve Henson]

  *) Deal with irritating shit to do with dependencies, in YAAHW (Yet Another
     Ad Hoc Way) - Makefile.ssls now all contain local dependencies, which
     can still be regenerated with "make depend".
     [Ben Laurie]

  *) Spelling mistake in C version of CAST-128.
     [Ben Laurie, reported by Jeremy Hylton <jeremy@cnri.reston.va.us>]

  *) Changes to the error generation code. The perl script err-code.pl 
     now reads in the old error codes and retains the old numbers, only
     adding new ones if necessary. It also only changes the .err files if new
     codes are added. The makefiles have been modified to only insert errors
     when needed (to avoid needlessly modifying header files). This is done
     by only inserting errors if the .err file is newer than the auto generated
     C file. To rebuild all the error codes from scratch (the old behaviour)
     either modify crypto/Makefile.ssl to pass the -regen flag to err_code.pl
     or delete all the .err files.
     [Steve Henson]

  *) CAST-128 was incorrectly implemented for short keys. The C version has
     been fixed, but is untested. The assembler versions are also fixed, but
     new assembler HAS NOT BEEN GENERATED FOR WIN32 - the Makefile needs fixing
     to regenerate it if needed.
     [Ben Laurie, reported (with fix for C version) by Jun-ichiro itojun
      Hagino <itojun@kame.net>]

  *) File was opened incorrectly in randfile.c.
     [Ulf Möller <ulf@fitug.de>]

  *) Beginning of support for GeneralizedTime. d2i, i2d, check and print
     functions. Also ASN1_TIME suite which is a CHOICE of UTCTime or
     GeneralizedTime. ASN1_TIME is the proper type used in certificates et
     al: it's just almost always a UTCTime. Note this patch adds new error
     codes so do a "make errors" if there are problems.
     [Steve Henson]

  *) Correct Linux 1 recognition in config.
     [Ulf Möller <ulf@fitug.de>]

  *) Remove pointless MD5 hash when using DSA keys in ca.
     [Anonymous <nobody@replay.com>]

  *) Generate an error if given an empty string as a cert directory. Also
     generate an error if handed NULL (previously returned 0 to indicate an
     error, but didn't set one).
     [Ben Laurie, reported by Anonymous <nobody@replay.com>]

  *) Add prototypes to SSL methods. Make SSL_write's buffer const, at last.
     [Ben Laurie]

  *) Fix the dummy function BN_ref_mod_exp() in rsaref.c to have the correct
     parameters. This was causing a warning which killed off the Win32 compile.
     [Steve Henson]

  *) Remove C++ style comments from crypto/bn/bn_local.h.
     [Neil Costigan <neil.costigan@celocom.com>]

  *) The function OBJ_txt2nid was broken. It was supposed to return a nid
     based on a text string, looking up short and long names and finally
     "dot" format. The "dot" format stuff didn't work. Added new function
     OBJ_txt2obj to do the same but return an ASN1_OBJECT and rewrote 
     OBJ_txt2nid to use it. OBJ_txt2obj can also return objects even if the
     OID is not part of the table.
     [Steve Henson]

  *) Add prototypes to X509 lookup/verify methods, fixing a bug in
     X509_LOOKUP_by_alias().
     [Ben Laurie]

  *) Sort openssl functions by name.
     [Ben Laurie]

  *) Get the gendsa program working (hopefully) and add it to app list. Remove
     encryption from sample DSA keys (in case anyone is interested the password
     was "1234").
     [Steve Henson]

  *) Make _all_ *_free functions accept a NULL pointer.
     [Frans Heymans <fheymans@isaserver.be>]

  *) If a DH key is generated in s3_srvr.c, don't blow it by trying to use
     NULL pointers.
     [Anonymous <nobody@replay.com>]

  *) s_server should send the CAfile as acceptable CAs, not its own cert.
     [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]

  *) Don't blow it for numeric -newkey arguments to apps/req.
     [Bodo Moeller <3moeller@informatik.uni-hamburg.de>]

  *) Temp key "for export" tests were wrong in s3_srvr.c.
     [Anonymous <nobody@replay.com>]

  *) Add prototype for temp key callback functions
     SSL_CTX_set_tmp_{rsa,dh}_callback().
     [Ben Laurie]

  *) Make DH_free() tolerate being passed a NULL pointer (like RSA_free() and
     DSA_free()). Make X509_PUBKEY_set() check for errors in d2i_PublicKey().
     [Steve Henson]

  *) X509_name_add_entry() freed the wrong thing after an error.
     [Arne Ansper <arne@ats.cyber.ee>]

  *) rsa_eay.c would attempt to free a NULL context.
     [Arne Ansper <arne@ats.cyber.ee>]

  *) BIO_s_socket() had a broken should_retry() on Windoze.
     [Arne Ansper <arne@ats.cyber.ee>]

  *) BIO_f_buffer() didn't pass on BIO_CTRL_FLUSH.
     [Arne Ansper <arne@ats.cyber.ee>]

  *) Make sure the already existing X509_STORE->depth variable is initialized
     in X509_STORE_new(), but document the fact that this variable is still
     unused in the certificate verification process.
     [Ralf S. Engelschall]

  *) Fix the various library and apps files to free up pkeys obtained from
     X509_PUBKEY_get() et al. Also allow x509.c to handle netscape extensions.
     [Steve Henson]

  *) Fix reference counting in X509_PUBKEY_get(). This makes
     demos/maurice/example2.c work, amongst others, probably.
     [Steve Henson and Ben Laurie]

  *) First cut of a cleanup for apps/. First the `ssleay' program is now named
     `openssl' and second, the shortcut symlinks for the `openssl <command>'
     are no longer created. This way we have a single and consistent command
     line interface `openssl <command>', similar to `cvs <command>'.
     [Ralf S. Engelschall, Paul Sutton and Ben Laurie]

  *) ca.c: move test for DSA keys inside #ifndef NO_DSA. Make pubkey
     BIT STRING wrapper always have zero unused bits.
     [Steve Henson]

  *) Add CA.pl, perl version of CA.sh, add extended key usage OID.
     [Steve Henson]

  *) Make the top-level INSTALL documentation easier to understand.
     [Paul Sutton]

  *) Makefiles updated to exit if an error occurs in a sub-directory
     make (including if user presses ^C) [Paul Sutton]

  *) Make Montgomery context stuff explicit in RSA data structure.
     [Ben Laurie]

  *) Fix build order of pem and err to allow for generated pem.h.
     [Ben Laurie]

  *) Fix renumbering bug in X509_NAME_delete_entry().
     [Ben Laurie]

  *) Enhanced the err-ins.pl script so it makes the error library number 
     global and can add a library name. This is needed for external ASN1 and
     other error libraries.
     [Steve Henson]

  *) Fixed sk_insert which never worked properly.
     [Steve Henson]

  *) Fix ASN1 macros so they can handle indefinite length construted 
     EXPLICIT tags. Some non standard certificates use these: they can now
     be read in.
     [Steve Henson]

  *) Merged the various old/obsolete SSLeay documentation files (doc/xxx.doc)
     into a single doc/ssleay.txt bundle. This way the information is still
     preserved but no longer messes up this directory. Now it's new room for
     the new set of documenation files.
     [Ralf S. Engelschall]

  *) SETs were incorrectly DER encoded. This was a major pain, because they
     shared code with SEQUENCEs, which aren't coded the same. This means that
     almost everything to do with SETs or SEQUENCEs has either changed name or
     number of arguments.
     [Ben Laurie, based on a partial fix by GP Jayan <gp@nsj.co.jp>]

  *) Fix test data to work with the above.
     [Ben Laurie]

  *) Fix the RSA header declarations that hid a bug I fixed in 0.9.0b but
     was already fixed by Eric for 0.9.1 it seems.
     [Ben Laurie - pointed out by Ulf Möller <ulf@fitug.de>]

  *) Autodetect FreeBSD3.
     [Ben Laurie]

  *) Fix various bugs in Configure. This affects the following platforms:
     nextstep
     ncr-scde
     unixware-2.0
     unixware-2.0-pentium
     sco5-cc.
     [Ben Laurie]

  *) Eliminate generated files from CVS. Reorder tests to regenerate files
     before they are needed.
     [Ben Laurie]

  *) Generate Makefile.ssl from Makefile.org (to keep CVS happy).
     [Ben Laurie]


 Changes between 0.9.1b and 0.9.1c  [23-Dec-1998]

  *) Added OPENSSL_VERSION_NUMBER to crypto/crypto.h and 
     changed SSLeay to OpenSSL in version strings.
     [Ralf S. Engelschall]
  
  *) Some fixups to the top-level documents.
     [Paul Sutton]

  *) Fixed the nasty bug where rsaref.h was not found under compile-time
     because the symlink to include/ was missing.
     [Ralf S. Engelschall]

  *) Incorporated the popular no-RSA/DSA-only patches 
     which allow to compile a RSA-free SSLeay.
     [Andrew Cooke / Interrader Ldt., Ralf S. Engelschall]

  *) Fixed nasty rehash problem under `make -f Makefile.ssl links'
     when "ssleay" is still not found.
     [Ralf S. Engelschall]

  *) Added more platforms to Configure: Cray T3E, HPUX 11, 
     [Ralf S. Engelschall, Beckmann <beckman@acl.lanl.gov>]

  *) Updated the README file.
     [Ralf S. Engelschall]

  *) Added various .cvsignore files in the CVS repository subdirs
     to make a "cvs update" really silent.
     [Ralf S. Engelschall]

  *) Recompiled the error-definition header files and added
     missing symbols to the Win32 linker tables.
     [Ralf S. Engelschall]

  *) Cleaned up the top-level documents;
     o new files: CHANGES and LICENSE
     o merged VERSION, HISTORY* and README* files a CHANGES.SSLeay 
     o merged COPYRIGHT into LICENSE
     o removed obsolete TODO file
     o renamed MICROSOFT to INSTALL.W32
     [Ralf S. Engelschall]

  *) Removed dummy files from the 0.9.1b source tree: 
     crypto/asn1/x crypto/bio/cd crypto/bio/fg crypto/bio/grep crypto/bio/vi
     crypto/bn/asm/......add.c crypto/bn/asm/a.out crypto/dsa/f crypto/md5/f
     crypto/pem/gmon.out crypto/perlasm/f crypto/pkcs7/build crypto/rsa/f
     crypto/sha/asm/f crypto/threads/f ms/zzz ssl/f ssl/f.mak test/f
     util/f.mak util/pl/f util/pl/f.mak crypto/bf/bf_locl.old apps/f
     [Ralf S. Engelschall]

  *) Added various platform portability fixes.
     [Mark J. Cox]

  *) The Genesis of the OpenSSL rpject:
     We start with the latest (unreleased) SSLeay version 0.9.1b which Eric A.
     Young and Tim J. Hudson created while they were working for C2Net until
     summer 1998.
     [The OpenSSL Project]
 

 Changes between 0.9.0b and 0.9.1b  [not released]

  *) Updated a few CA certificates under certs/
     [Eric A. Young]

  *) Changed some BIGNUM api stuff.
     [Eric A. Young]

  *) Various platform ports: OpenBSD, Ultrix, IRIX 64bit, NetBSD, 
     DGUX x86, Linux Alpha, etc.
     [Eric A. Young]

  *) New COMP library [crypto/comp/] for SSL Record Layer Compression: 
     RLE (dummy implemented) and ZLIB (really implemented when ZLIB is
     available).
     [Eric A. Young]

  *) Add -strparse option to asn1pars program which parses nested 
     binary structures 
     [Dr Stephen Henson <shenson@bigfoot.com>]

  *) Added "oid_file" to ssleay.cnf for "ca" and "req" programs.
     [Eric A. Young]

  *) DSA fix for "ca" program.
     [Eric A. Young]

  *) Added "-genkey" option to "dsaparam" program.
     [Eric A. Young]

  *) Added RIPE MD160 (rmd160) message digest.
     [Eric A. Young]

  *) Added -a (all) option to "ssleay version" command.
     [Eric A. Young]

  *) Added PLATFORM define which is the id given to Configure.
     [Eric A. Young]

  *) Added MemCheck_XXXX functions to crypto/mem.c for memory checking.
     [Eric A. Young]

  *) Extended the ASN.1 parser routines.
     [Eric A. Young]

  *) Extended BIO routines to support REUSEADDR, seek, tell, etc.
     [Eric A. Young]

  *) Added a BN_CTX to the BN library.
     [Eric A. Young]

  *) Fixed the weak key values in DES library
     [Eric A. Young]

  *) Changed API in EVP library for cipher aliases.
     [Eric A. Young]

  *) Added support for RC2/64bit cipher.
     [Eric A. Young]

  *) Converted the lhash library to the crypto/mem.c functions.
     [Eric A. Young]

  *) Added more recognized ASN.1 object ids.
     [Eric A. Young]

  *) Added more RSA padding checks for SSL/TLS.
     [Eric A. Young]

  *) Added BIO proxy/filter functionality.
     [Eric A. Young]

  *) Added extra_certs to SSL_CTX which can be used
     send extra CA certificates to the client in the CA cert chain sending
     process. It can be configured with SSL_CTX_add_extra_chain_cert().
     [Eric A. Young]

  *) Now Fortezza is denied in the authentication phase because
     this is key exchange mechanism is not supported by SSLeay at all.
     [Eric A. Young]

  *) Additional PKCS1 checks.
     [Eric A. Young]

  *) Support the string "TLSv1" for all TLS v1 ciphers.
     [Eric A. Young]

  *) Added function SSL_get_ex_data_X509_STORE_CTX_idx() which gives the
     ex_data index of the SSL context in the X509_STORE_CTX ex_data.
     [Eric A. Young]

  *) Fixed a few memory leaks.
     [Eric A. Young]

  *) Fixed various code and comment typos.
     [Eric A. Young]

  *) A minor bug in ssl/s3_clnt.c where there would always be 4 0 
     bytes sent in the client random.
     [Edward Bishop <ebishop@spyglass.com>]