I have a project page at Google Code:
http://code.google.com/p/delphidec/

Give me some weeks, hopefully I’ll have finished a new release next month. Currently I don’t know if my idea is going to work, so don’t expect anything – but please stay tuned

Cheers,
Arvid

P.S.: Sorry don’t have time to look into old errors. Looks like something related to ver.inc, best to check which conditional compilation is correct for your delphi.

TDECHashClass does not contain a member ‘Register’ at line 444

Yes, you can build them directly from the sources from openssl.org.
That’s what I do, too.

Just that we pay attention which c/asm compiler is used to provide best support for all platforms (i.e. Win 2000+) without any dependencies in contrast to the famous SlProWeb OpenSSL dlls.

Special Indy adjustments are not needed anymore, that’s valid since Indy 10.5.5 (= Delphi 2009) and newer.

Cheers,
Arvid

Thanks.

We want to add a LongWord value multiplied by 8 to a large 256Bit unsigned integer. We multiply it by 8 because we want to count the Bits computed out of Bytes. Some Hash functions need to include the processed length of data streams in Bits into they computation. I had the choice to add the bytes processed into a large integer and finally multiply this large integer by 8, or to do this directly in this addition function. Later is faster and don’t need a separate multiply function to implement.

The mistake I made is in SHR,25 and must be SHR ECX,29.

ECX:EDX contains first our parameter Add and must be multiplied by 8. Thus we load EDX with LEA EDX,[EDX*8] = EDX *8. This lost the most three bits. We had loaded prior ECX with parameter Add and have to divide it with 2^32/8 = 2^32/2^3 = 2^(32-3) = SHR ECX,29 to correct these lost three most significant bits.

Simply change 25 to 29.

A pascal version of this code is complicated because we can’t access to the CPUs Carry flag in Pascal. Thats used in the assembler version to get the right mathematical result. In i386 assembler there is easily the possibility to chain multiple additions/subtractions of each 32bit in size to get a addition of arbitrary integer sizes, like this function does.

Best regards, Hagen

let me check that, I’ll reply you privately…

Cheers
Arvid

Define “later this year”

We use DEC in our product and are now starting to port it to x64.
That means we need DEC compiling and running on Delphi XE2 64bit mode. If you didn’t have time to port it yet, we will do it (at least these parts we need), but in this case it would be helpful if we could base on the latest code. Therefore I’d kindly ask you if it is possible to use/fill the SVN repository on Google Code with the latest code?

Thank you and best regards
Andreas