Hello, You are right... There were lot of try and fail play around the headers and I wanted to submit the patch before the sunset - therefore I haven't checked it over.
1. _XOPEN_SOURCE_EXTENDED issue. Yes, it would be good to have it defined for all VMS systems, but the main reason was the struct timeval that is in the time.h ... depending on that define. But I have found resource.h an include file made to solve issues around this. 2. It is a copy paste mistake, because of the reason explained above :( Part: + #if defined(OPENSSL_SYS_WIN32) || defined(OPENSSL_SYS_VMS) + #include <sys/timeb.h> + #endif Should not be there - please ignore it. Regards, Z -----Original Message----- From: Richard Levitte [mailto:rich...@levitte.org] Sent: den 17 augusti 2009 11:10 To: openssl-dev@openssl.org; Arpadffy Zoltan Subject: Re: [PATCH] OpenSSL 1.0.0 beta3 release v. VMS In message <839c820b5c926b4b89713b3a6ed68d2aae4...@sgstmail.scigames.at> on Fri, 14 Aug 2009 19:15:12 +0200, "Arpadffy Zoltan" <zoltan.arpad...@scientificgames.se> said: Zoltan.Arpadffy> Mandatory information: Zoltan.Arpadffy> Request type: PATCH Zoltan.Arpadffy> OS and arch: OpenVMS ALPHA, IA64 Zoltan.Arpadffy> OpenSSL version: 1.0.0 beta 3 I've looked through it, and applied almost all of it. There are just two things I wonder about, and haven't applied yet: In makevms.com: *************** $! *** 342,347 **** --- 343,349 ---- $ WRITE H_FILE "#undef OPENSSL_EXPORT_VAR_AS_FUNCTION" $ WRITE H_FILE "#define OPENSSL_EXPORT_VAR_AS_FUNCTION" $! + $ WRITE H_FILE "#define _XOPEN_SOURCE_EXTENDED" $! End $! $ ENDIF This would only define that logical for VAX. Is it really not needed for Alpha and ia64? Please understand me, adding that logical there is probably a very good idea... I've been conservative and only added it on file-per-file basis, and that's probably not the wisest. *** openssl-100-beta3/SSL/DTLS1.H Wed Jun 17 13:38:26 2009 --- openssl-100-beta3-work/SSL/DTLS1.H Fri Aug 14 09:40:39 2009 *************** *** 62,67 **** --- 62,74 ---- #include <openssl/buffer.h> #include <openssl/pqueue.h> + #if defined(OPENSSL_SYS_WIN32) || defined(OPENSSL_SYS_VMS) + #include <sys/timeb.h> + #endif + #ifdef OPENSSL_SYS_VMS + #include <resource.h> + #include <sys/timeb.h> + #endif #ifdef OPENSSL_SYS_WIN32 /* Needed for struct timeval */ #include <winsock.h> This seems odd to me. Or rather, the first added conditional. I doubt Win32 needs sys/timeb.h, as it uses a different mechanism (as far as I understand, I'm not the Windows hacker here ;-)), and it seems redundant to include sys/timeb.h twice for VMS. So far as I can see, only the second added conditional makes sense... Cheers, Richard -- Richard Levitte rich...@levitte.org http://richard.levitte.org/ "Life is a tremendous celebration - and I'm invited!" -- from a friend's blog, translated from Swedish ______________________________________________________________________ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager majord...@openssl.org
