Comments on: M2Crypto Build Wrapper for Fedora Core -based Distributions http://www.heikkitoivonen.net/blog/2009/02/09/m2crypto-build-wrapper-for-fedora-core-based-distributions/ A Finn in Silicon Valley - Adventures in Technology Wed, 17 Aug 2011 08:58:27 +0000 hourly 1 http://wordpress.org/?v=3.0.5 By: Andrew http://www.heikkitoivonen.net/blog/2009/02/09/m2crypto-build-wrapper-for-fedora-core-based-distributions/comment-page-1/#comment-12299 Andrew Mon, 15 Jun 2009 20:12:21 +0000 http://www.heikkitoivonen.net/blog/?p=694#comment-12299 I had to modify the script slightly, so arch is set to `uname -i` (i386) instead of `uname -m` (i686). This is on FC8. I had to modify the script slightly, so arch is set to `uname -i` (i386) instead of `uname -m` (i686). This is on FC8.

]]> By: Miloslav Trmač http://www.heikkitoivonen.net/blog/2009/02/09/m2crypto-build-wrapper-for-fedora-core-based-distributions/comment-page-1/#comment-8095 Miloslav Trmač Thu, 12 Mar 2009 09:50:45 +0000 http://www.heikkitoivonen.net/blog/?p=694#comment-8095 Doing it this way does not require any modification of the build process of applications that use openssl. M2Crypto is an exception because SWIG does not (or should not) follow #includes recursively, which breaks the platform auto-detection header file. Doing it this way does not require any modification of the build process of applications that use openssl.

M2Crypto is an exception because SWIG does not (or should not) follow #includes recursively, which breaks the platform auto-detection header file.

]]> By: Heikki Toivonen http://www.heikkitoivonen.net/blog/2009/02/09/m2crypto-build-wrapper-for-fedora-core-based-distributions/comment-page-1/#comment-8070 Heikki Toivonen Wed, 11 Mar 2009 03:42:52 +0000 http://www.heikkitoivonen.net/blog/?p=694#comment-8070 Hmm, I still don't see why this couldn't be solved with defines during build or something like that. Of course, I haven't really thought through all issues so maybe there really was no better way. But thanks for the explanation. Hmm, I still don’t see why this couldn’t be solved with defines during build or something like that. Of course, I haven’t really thought through all issues so maybe there really was no better way. But thanks for the explanation.

]]> By: Miloslav Trmač http://www.heikkitoivonen.net/blog/2009/02/09/m2crypto-build-wrapper-for-fedora-core-based-distributions/comment-page-1/#comment-8054 Miloslav Trmač Tue, 10 Mar 2009 14:17:21 +0000 http://www.heikkitoivonen.net/blog/?p=694#comment-8054 The opensslconf.h changes are necessary for multilib systems - to allow installing both i386 and x86_64 headers and development libraries on one machine, and building both 32-bit and 64-bit programs using those headers. If opensslconf.h contained data about the 64-bit mode, building 32-bit programs wouldn't be possible. The opensslconf.h changes are necessary for multilib systems – to allow installing both i386 and x86_64 headers and development libraries on one machine, and building both 32-bit and 64-bit programs using those headers. If opensslconf.h contained data about the 64-bit mode, building 32-bit programs wouldn’t be possible.

]]>