This is off the top of my head, didn't check the sources to make sure, so validity is 'mostly sure but not entirely'.
Quite a few bits 'n pieces in OpenSSL are 'template' based. (several crypto methods, a lot of ASN.1 work, an object stack, ...) Think about templates (wider scope than just the C++ flavor) and then consider how you would implement a mechanism like that in 'C': yep, through preprocessor work. That's what's going on (be reminded about the preprocessor ## operator at this point: it can be used to 'construct' variable, function, etc. names from multiple parts.) If you grep the OpenSSL source tree you'll find several macros with names like DECLARE_<something> and IMPLEMENT_<something>, etc. and those buggers perform the 'magic' of generating implementations for functions either declared directly or through generation by a DECLARE_...() macro in a header file. IIRC it's BLOCK_CIPHER_<something> macros you want to look for; should most probably hang around in a header file in crypto/evp. Grep for the preproc concatenation operator (##) if you're not getting results otherwise; that grep should lead you to any macros which might generate functions like the ones you mention. Your linker problem is, to me, a hint that you may have simply missed to include one or two of those evp .c files in your cmake build process; those .c files will make sure the functions you are missing are indeed generated through the preproc and those BLOCK_CIPHER_xyz macros (assuming I have the 'BLOCK_CIPHER' bit right, that is. Don't take my word for it; check the header files in crypto/evp and possibly crypto/des for any '##' and percolate up from there. No voodoo/dead chicken required as long as you stick with the ## basic move. ;-) On Tue, Jul 6, 2010 at 2:02 PM, Deckers, Rob <rob.deck...@oce.com> wrote: > Thanks, > > I am starting to understand the structure. > I now made the same structure in my CMake files. > > 11 engines and the modules crypto and ssl. > > But the Crypto module has a linker problem. > > > 1>------ Build started: Project: openssl.crypto, Configuration: Debug Win32 > ------ > 1>Linking... > 1> Creating library > Y:\solutions\si-refinfrablue\proglibs\OpenSSL\crypto\Debug\openssl.crypto.lib > and object > Y:\solutions\si-refinfrablue\proglibs\OpenSSL\crypto\Debug\openssl.crypto.exp > 1>c_allc.obj : error LNK2019: unresolved external symbol _EVP_des_ecb > referenced in function _OpenSSL_add_all_ciphers > 1>c_allc.obj : error LNK2019: unresolved external symbol _EVP_des_cbc > referenced in function _OpenSSL_add_all_ciphers > 1>c_allc.obj : error LNK2019: unresolved external symbol _EVP_des_ofb > referenced in function _OpenSSL_add_all_ciphers > 1>c_allc.obj : error LNK2019: unresolved external symbol _EVP_des_cfb8 > referenced in function _OpenSSL_add_all_ciphers > 1>c_allc.obj : error LNK2019: unresolved external symbol _EVP_des_cfb1 > referenced in function _OpenSSL_add_all_ciphers > 1>c_allc.obj : error LNK2019: unresolved external symbol _EVP_des_cfb64 > referenced in function _OpenSSL_add_all_ciphers > 1>e_old.obj : error LNK2001: unresolved external symbol _EVP_des_cfb64 > 1>Y:\solutions\si-refinfrablue\proglibs\OpenSSL\crypto\Debug\openssl.crypto.dll > : fatal error LNK1120: 6 unresolved externals > 1>Build log was saved at > file://y:\solutions\si-refinfrablue\proglibs\OpenSSL\crypto\openssl.crypto.dir\Debug\BuildLog.htm > 1>openssl.crypto - 8 error(s), 0 warning(s) > ========== Build: 0 succeeded, 1 failed, 1 up-to-date, 0 skipped ========== > When I search the code for "EVP_des_ecb", I find it in evp.h > 00659: const EVP_CIPHER *EVP_des_ecb(void); > 00660: const EVP_CIPHER *EVP_des_ede(void); > 00661: const EVP_CIPHER *EVP_des_ede3(void); > 00662: const EVP_CIPHER *EVP_des_ede_ecb(void); > 00663: const EVP_CIPHER *EVP_des_ede3_ecb(void); > 00664: const EVP_CIPHER *EVP_des_cfb64(void); > and c_allc.c. > 00092: EVP_add_cipher(EVP_des_ecb()); > 00093: EVP_add_cipher(EVP_des_ede()); > 00094: EVP_add_cipher(EVP_des_ede3()); > But I can't find an implementation for these functions and the 'normal' > OpenSSL build evironment doesn't seem to have problems with these functions. > > > Rob Deckers > > This message and attachment(s) are intended solely for use by the addressee > and may contain information that is privileged, confidential or otherwise > exempt from disclosure under applicable law. If you are not the intended > recipient or agent thereof responsible for delivering this message to the > intended recipient, you are hereby notified that any dissemination, > distribution or copying of this communication is strictly prohibited. If you > have received this communication in error, please notify the sender > immediately by telephone and with a 'reply' message. Thank you for your > co-operation. > -- Met vriendelijke groeten / Best regards, Ger Hobbelt -------------------------------------------------- web: http://www.hobbelt.com/ http://www.hebbut.net/ mail: g...@hobbelt.com mobile: +31-6-11 120 978 --------------------------------------------------
