update optional package openssl to version 1.0.1a

[mariah]

Upgrade the optional package openssl to version 1.0.1a which was released on 19 April 2012.

Version with tarball removed:

spkg:  http://wstein.org/home/wstein/patches/openssl-1.0.1a.spkg

[was]

Can you change the permissions:

wstein@sage:/scratch/wstein/ref/sage-5.0.beta15-boxen-x86_64-Linux$ wget  http://boxen.math.washington.edu/home/mariah/spkg/openssl-1.0.1a.spkg
--13:51:00--  http://boxen.math.washington.edu/home/mariah/spkg/openssl-1.0.1a.spkg
           => `openssl-1.0.1a.spkg'
Resolving boxen.math.washington.edu... 128.208.160.197
Connecting to boxen.math.washington.edu|128.208.160.197|:80... connected.
HTTP request sent, awaiting response... 403 Forbidden
13:51:00 ERROR 403: Forbidden.

[mariah]

Done. Apologies.

[was]

REPORT:

(this is long, but there is really only one trivial change I want made.)

1. Delete this file -- why is it there?

   -rw-r-----  1 wstein wstein  21M 2012-04-19 05:20 openssl-1.0.1a.tar
   

2. It builds fine on sage.math (64-bit ubuntu).

3. It fails to build on OS X 10.7.3 after building sage-5.0.pre15 out of the box:

   gcc -I. -I.. -I../include  -fPIC -fno-common -DOPENSSL_PIC -DOPENSSL_THREADS -D_REENTRANT -DDSO_DLFCN -DHAVE_DLFCN_H -arch i386 -O3 -fomit-frame-pointer -DL_ENDIAN -DOPENSSL_BN_ASM_PART_WORDS -DOPENSSL_IA32_SSE2 -DOPENSSL_BN_ASM_MONT -DOPENSSL_BN_ASM_GF2m -DSHA1_ASM -DSHA256_ASM -DSHA512_ASM -DMD5_ASM -DRMD160_ASM -DAE
   S_ASM -DVPAES_ASM -DWHIRLPOOL_ASM -DGHASH_ASM   -c -o mem_dbg.o mem_dbg.c
   gcc: error: i386: No such file or directory
   gcc: error: unrecognized option ‘-arch’
   gcc: error: i386: No such file or directory
   gcc: error: unrecognized option ‘-arch’
   make[1]: *** [cryptlib.o] Error 1
   make[1]: *** Waiting for unfinished jobs....
   make[1]: *** [mem.o] Error 1
   gcc: error: i386: No such file or directory
   gcc: error: unrecognized option ‘-arch’
   make[1]: *** [mem_dbg.o] Error 1
   make: *** [build_crypto] Error 1
   Error building openssl.
   
   real	0m3.738s
   user	0m2.527s
   sys	0m0.910s
   ************************************************************************
   Error installing package openssl-1.0.1a
   

This looks like some OS version detection code is wrong? Note gcc: error: unrecognized option ‘-arch’. Exactly the same problem happens with the openssl-1.0.0.p0 that we currently have in the optional package repo.

4. On Ubuntu 64-bit (sage.math), I tried rebuilding Python after building the new openssl, and that worked fine and did include openssl support.

5. I ran all doctests on ubuntu 64-bit and that worked fine. This isn't surprising, since I think no doctests use openssl.

6. On OS X 10.6, I tried this:

   bsd:sage-5.0.beta8 wstein$ ./sage -i  http://boxen.math.washington.edu/home/mariah/spkg/openssl-1.0.1a.spkg
   

and it worked fine.

7. It would be nice if there was an ".hgignore" file since

   $ hg status
   ? openssl-1.0.1a.spkg
   ? openssl-1.0.1a/SPKG.txt
   ? openssl-1.0.1a/openssl-1.0.1a.tar
   ? openssl-1.0.1a/src/ACKNOWLEDGMENTS
   ? openssl-1.0.1a/src/CHANGES
   ? openssl-1.0.1a/src/CHANGES.SSLeay
   ? openssl-1.0.1a/src/Configure
   ...
   

SUMMARY: The OS X 10.7.3 issue looks serious, but should not hold this back, because it is already in the current openssl optional package; it should be considered a new bug on a separate ticket. Otherwise, we just need to delete the upstream tarball which is there for no reason. Since this is an optional spkg, and we still don't officially support OS X 10.7.3, I don't consider this a total blocker for a positive review.

THUS: Positive review contingent on deleting that one tar file.

[jdemeyer]

Which is the correct spkg?

[jdemeyer]

*ping*

please clarify: why are there two spkg's listed?

[mariah]

The spkg posted by William supercedes my spkg. (William's spkg removes a tar file that I mistakenly left in the spkg.)

I have removed my spkg from the Description.

I am setting this back to positive review because William gave it a positive review.

[schilly]

spkg is on the webserver+mirros now.

[jhpalmieri]

See #13126 for a follow-up.