Excerpts from Matthias Klose's message of 2012-06-20 02:46:19 -0700: > On 19.06.2012 17:54, Nicholas Bamber wrote: > >>>>> 1.) compile against gcc-4.5 and g++-4.5 > >>>>> 2.) set the magic TAOCRYPT_DISABLE_X86ASM thingy causing SSL connections > >>>>> on those platforms to be slower. > >>>>> 3.) compile against gcc-4.4 and g++-4.4 > > > > > > Matthias, > > Any opinion on which short term hack we should adopt? Would you like to > > ahve a look at the code which is causing us issues? > > this is inline assembler, I won't look at it. So best disable it if nobody > wants to maintain it (and the code has comments like "won't work with compiler > xy at opt level z"). If you still want to look at it, find out which function > with the inline assembler causes the unexpected behaviour (modify the code to > build just half of the functions with asm statements, then bisect until you > get > down to one function). > > As a side note, why isn't the package built using the system openssl? It looks > like everybody except Debian/Ubuntu does do this. >
Its been explained to me like this: * libmysqlclient is GPL, with an exception given to things that are not GPL to link to it: http://www.mysql.com/about/legal/licensing/foss-exception/ * OpenSSL is given an exception, despite being GPL incompatible * However, now anything that links to *libmysqlclient* must also explicitly give OpenSSL an exception since they are effectively transitively accepting the OpenSSL license, which is not compatible. With almost 150 reverse depends on libmysqlclient, it would be a lot of work to make sure all of those projects gave OpenSSL an exception. -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org