1 2001-09-07 Gisle Aas <gisle@ActiveState.com>
5 Sync up with the bleadperl version:
6 - use SvPVbyte() if avaiable
7 - fixes to make the code 'gcc -Wall'-clean
11 2001-08-27 Gisle Aas <gisle@ActiveState.com>
15 Avoid exit() in Makefile.PL and bleedperls redefinition of printf
16 in the alignment test program.
17 Patch by Doug MacEachern <dougm@covalent.net>.
21 2001-07-18 Gisle Aas <gisle@ActiveState.com>
25 Try to warn if the functional interface is used as methods,
26 i.e. Digest::MD5->md5_hex("foo") will make noise if -w is
29 Document the missing padding for the base64 digests.
31 If both XS bootstrap and locating Digest::Perl::MD5 fails
32 reraise the original XS bootstrap exception.
36 2001-03-13 Gisle Aas <gisle@ActiveState.com>
40 Moved all other Digest:: modules out of the Digest-MD5 dist.
44 2000-09-18 Gisle Aas <gisle@ActiveState.com>
48 Avoid pointer cast warning for machines with bigger ints
49 than pointers. Patch by Robin Barker <rmb1@cise.npl.co.uk>.
53 2000-08-19 Gisle Aas <gisle@ActiveState.com>
57 The fallback code introduced in 2.10 did only work for
58 perl-5.6.0. It should now for for perl5.004 and 5.005
59 as well. Patch by Ville Skyttä <ville@office.popsystems.com>.
63 2000-08-18 Gisle Aas <gisle@ActiveState.com>
67 Digest::MD5 will now try to fallback to the pure perl
68 implementation of Digest::Perl::MD5 if bootstrap fails.
70 Added a bit internal paranoia about casting the IV
71 in the Digest::MD5 object to the MD5_CTX* pointer.
75 1999-08-06 Gisle Aas <gisle@aas.no>
83 1999-07-28 Gisle Aas <gisle@aas.no>
87 The addfile() methods could trigger a core dump when passed
88 a filehandle that had failed to open.
92 1999-04-26 Gisle Aas <gisle@aas.no>
96 The Digest::SHA1 module failed on some 64-bit systems, because I
97 assumed there was a correspondence between the U32 size and
98 BYTEORDER. This version use 'unsigned long' as Uwe's original
101 The module should now work better when liked statically with perl,
102 because we now use a safer module-loaded test in Digest.pm.
104 Assume we know the outcome of the alignment test on VMS. Patch by
105 Chuck Lane <lane@duphy4.physics.drexel.edu>
109 1999-03-26 Gisle Aas <gisle@aas.no>
113 Avoid LONG and BYTE types in SHA.xs as they was in conflict
114 with similar definitions in <winnt.h>.
116 Patch by Marko Asplund <aspa@hip.fi> to make the the alignment
117 test program link successfully with sfio-perl.
119 Fixed a typo in MD5.xs that might have affected 64-bit systems.
120 Spotted by Nick Ing-Simmons
124 1999-03-15 Gisle Aas <gisle@aas.no>
128 Included Digest::SHA1 based on Uwe Hollerbach's SHA module.
132 1999-03-05 Gisle Aas <gisle@aas.no>
136 Avoid the -o option when compiling alignment test program
137 for Win32 as suggested by Gurusamy Sarathy.
139 DEC Compiler bug workaround. Contributed by D Roland Walker
140 <walker@ncbi.nlm.nih.gov>
142 Having references to a local variable called "na" was not
143 very safe either. Some older versions of Perl can apparently
144 macroize this into something completely different.
148 1999-02-27 Gisle Aas <gisle@aas.no>
152 Patch from Christopher J. Madsen <chris_madsen@geocities.com> that
153 should help getting the u32align test program to compile with
154 Visual C++ 5 on Windows NT.
156 Got rid of references to PL_na.
160 1999-01-31 Gisle Aas <gisle@aas.no>
164 Added a hints file as workaround for an IRIX compiler bug.
165 Contributed by D Roland Walker <walker@ncbi.nlm.nih.gov>.
167 Note that the rfc2202 test can still fail on some DEC Alpha,
168 because of a compiler bug that affects the perl 'x' operator.
169 The Digest:: modules should work and be safe to install anyway.
173 1998-12-18 Gisle Aas <aas@sn.no>
177 Some casts and tweaks to make picky compilers more happy.
181 1998-11-04 Gisle Aas <aas@sn.no>
185 Taken out Digest::SHA1 as this module will be provided from Uwe
188 Some tweaks to MD2.xs and MD5.xs since "na" disappeared in
193 1998-10-30 Gisle Aas <aas@sn.no>
197 The 1.99_59 release introduced compilation problems for big-endian
198 systems with free U32 alignment. Bug reported, and fix suggested
199 by Paul J. Schinder <schinder@pobox.com>.
203 1998-10-28 Gisle Aas <aas@sn.no>
207 Makefile.PL will run a test program to find out if U32 values can
208 be aligned anywhere. This hopefully cures the core dumps reported
209 on Solaris and other big endian systems. Thanks to Graham Barr for
214 1998-10-28 Gisle Aas <aas@sn.no>
218 Should be very close to a 2.00 release now. Need some success
219 reports from people running on big-endian machines first I think.
221 Added a Digest::MD2 implementation.
223 Wrote Digest.pm documentation. This define the interface that all
224 Digest:: modules should provide.
226 Avoided some code duplication in MD5.xs
228 Fixed typo, that prevented Digest::SHA1::sha1_base64() from working.
232 1998-10-27 Gisle Aas <aas@sn.no>
236 Rewritten most of the MD5 C code to make it real fast (especially
237 on little-endian machines without alignment restrictions for U32).
238 Compared to MD5-1.7 we can process files 4 times as fast and we
239 digest small stuff in memory 7 times faster. I came to these
240 conclusions after these tests (gcc -O2, i586, Linux):
242 First tested calculation of the digest of a 31 MB file, using
243 perl -le 'print Digest::MD5->new->addfile(*STDIN)->hexdigest'
247 Digest::MD5-1.99_57: 5.23s
248 md5sum (GNU textutils): 4.90s
250 As you can see, we do nearly as good as the md5sum program. I
251 think the reason we don't beat md5sum is that perl always insist on
252 loading extra modules like Config.pm, Carp.pm, strict.pm, vars.pm,
253 AutoLoader.pm and DynaLoader.pm. When I simply wrapped the MD5.xs
254 hasher code in a C program I managed to process the file in 4.68s.
256 Then we calculated the digest of the same 6 byte sting, 20000
260 Digest::MD5-1.99_57: 1.68s
262 Digest::MD5 benefit from making this into a plain procedure call
263 instead of a static method call.
266 Other changes in this release are:
270 Internal MD5.xs cleanup.
272 $md5->digest will automatically reset now.
274 Digest::HMAC methods add() and addfile() did not return the
277 Added Digest.pm loading module. I am not sure this is a good idea.
279 Added Digest::SHA1 and Digest::HMAC_SHA1 module. The Digest::SHA1
280 module is just a wrapper around SHA.pm. I hope to get the author
281 of SHA.pm to move his module to the Digest:: category.
285 1998-10-25 Gisle Aas <aas@sn.no>
289 Fix memcpy_byteswap() function in MD5.xs. Must be careful with
290 htovl() as it might evaluate its arguments more than once.
294 1998-10-25 Gisle Aas <aas@sn.no>
298 Grahams HMAC_MD5.pm splitted into two modules. Digest::HMAC and
299 Digest::HMAC_MD5. Also provide functional interface. Documentation
302 Included RFC 2202 based test for HMAC-MD5.
306 1998-10-24 Gisle Aas <aas@sn.no>
310 Included HMAC_MD5.pm, contributed by Graham Barr <gbarr@ti.com>.
312 I have a hard time to make up my mind :-) md5_bin() renamed back
313 to md5(). Functions are not exported by default any more.
315 Try to Encode/Decode with memcpy_byteswap for 32-bit big-endian
320 1998-10-23 Gisle Aas <aas@sn.no>
324 Renamed core module as Digest::MD5. Leave a MD5.pm stub for
327 The md5() function renamed as md5_bin().
329 The constructor, Digest::MD5->new, no longer takes any extra
332 Added some new tests.
334 Updated the documentation.
336 $md5->b64digest implemented with same base64 encoder as md5_base64.
340 1998-10-23 Gisle Aas <aas@sn.no>
344 Patch from Graham Barr which make it work for big-endian machines
349 1998-10-22 Gisle Aas <aas@sn.no>
353 The MD5 class is now subclassable.
355 The add() and addfile() methods now return $self.
357 The reset() method is just an alias for new().
359 The constructor (MD5->new) now takes optional arguments which are
360 automatically added. It means that we can now write:
362 MD5->new($data)->hexdigest;
364 New $md5->b64digest method.
366 New functions that are exported on request: md5, md5_hex, md5_base64
370 Barely started to update the documentation.
374 1998-10-22 Gisle Aas <aas@sn.no>
378 Much better performance (more than twice as fast now). Mostly
379 because we use Copy/Zero instead of the original MD5_memcpy and
380 MD5_memset functions.
382 The addfile() and hexdigest() methods are now XS implemented.
384 All RSA functions now included in MD5.xs and made static.
386 Use perl's Copy/Zero.
388 Random cleanup, simplifications and reformatting.
389 Merged things better with the perl configuration.
393 Neil Winton's versions below:
396 *** 96/06/20 Version 1.7
398 MD5 is now completely 64-bit clean (I hope). The basic MD5 code uses
399 32-bit quantities and requires a typedef UINT4 to be defined in
400 global.h. Perl configuration data (the value of BYTEORDER) is used to
401 determine if unsigned longs have 4 or 8 bytes. On 64-bit platforms (eg
402 DEC Alpha) then it assumes that "unsigned int" will be a 32-bit type.
403 If this is incorrect then adding -DUINT4_IS_LONG to the DEFINES line in
404 Makefile.PL will override this.
406 On some machines (at least Cray that I know of) there is no 32-bit
407 integer type. In this case defining TRUNCATE_UINT4 (which is done
408 automatically for a Cray) will ensure that 64-bit values are masked
409 down to 32 bits. I have done my best to test this but without easy
410 access to a true 64-bit machine I can not totally guarantee it (unless
411 anyone wants to lend me a spare Cray :-)
413 There is one remaining limitation for 64-bit enabled processors. The
414 amount of data passed to any single call to the underlying MD5
415 routines is limited to (2^32 - 1) bytes -- that's 4 gigabytes. I'm
416 sorry if that's a real problem for you ...
418 And finally, a minor complilation warning (unsigned char * used with
419 function having char * prototype) has also been eliminated.
421 *** 96/04/09 Version 1.6
423 Re-generated module framework using h2xs to pick up the latest module
424 conventions for versions etc. You can now say "use MD5 1.6;" and things
425 should work correctly. MD5.pod has been integrated into MD5.pm and
426 CHANGES renamed to Changes. There is a fairly comprehensive test.pl
427 which can be invoked via "make test". There are no functional changes
428 to the MD5 routines themselves.
430 *** 96/03/14 Version 1.5.3
432 Fixed addfile method to accept type-glob references for the file-handle
433 (eg \*STDOUT). This is more consistent with other routines and is now the
434 recommended way of passing file-handles. The documentation now gives more
435 examples as to how the routines might be used.
437 *** 96/03/12 Version 1.5.2
439 Minor fixes from Christopher J Madsen <madsen@computek.net> to provide
440 support for building on OS/2 (and to work arround a perl -w bug).
442 Remove warning about possible difference between add('foo', 'bar') and
443 add('foobar'). This is not true (it may have been true in the earliest
444 version of the module but is no longer the case).
446 *** 96/03/08 Version 1.5.1
448 Add CHANGES file to make it easier for people to figure out what has
449 been going on. (Meant to do this as part of 1.5)
451 *** 96/03/05 Version 1.5
453 Add hash() and hexhash() methods at the suggestion/request of Gary
454 Howland <gary@kampai.euronet.nl> before inclusion in a wider library
455 of cryptography modules.
457 *** 96/02/27 Version 1.4
459 Finally fixed the pesky Solaris dynamic loading bug. All kudos to Ken
460 Pizzini <kenp@spry.com>!
462 *** 95/11/29 Version 1.3.1
464 Add explanations of current known problems.
466 *** 95/06/02 Version 1.3
468 Fix problems with scope resolution in addfile() reported by
469 Jean-Claude Giese <Jean-Claude.Giese@loria.fr>. Basically ARGV is
470 always implicitly in package main while other filehandles aren't.
472 *** 95/05/23 Version 1.2.1
474 [Changes pre 1.2.1 not recorded]