Problem building Cryptopp562 on Mac OSX10.8 using GCC4.8.2

404 views
Skip to first unread message

Tim Burgess

unread,
Dec 6, 2013, 11:21:49 AM12/6/13
to cryptop...@googlegroups.com
Hi,
 
I downloaded the latest build using SVN then got the following output when I attempted a build:
 

macbook_pro:cryptopp tim$ make static

c++ -DNDEBUG -g -O2 -arch x86_64 -arch i386 -DCRYPTOPP_DISABLE_ASM -pipe -c 3way.cpp

In file included from /opt/local/include/gcc48/c++/bits/stl_algo.h:59:0,

                 from /opt/local/include/gcc48/c++/algorithm:62,

                 from stdcpp.h:18,

                 from cryptlib.h:83,

                 from seckey.h:8,

                 from 3way.h:7,

                 from 3way.cpp:5:

/opt/local/include/gcc48/c++/cstdlib:178:10: error: expected unqualified-id before '__int128'

   inline __int128

          ^

make: *** [3way.o] Error 1

 
GCC4.8.1 is my default c/c++ compiler and lives in /opt/local (I installed it from MacPorts). I'm fairly new to developing on the Mac, so can anybody please advise me of what I might try to build the library?
 
Many thanks in advance.
 
Tim

Jeffrey Walton

unread,
Dec 10, 2013, 7:54:32 PM12/10/13
to cryptop...@googlegroups.com

-arch i386 is the problem. See http://stackoverflow.com/questions/13148463/compiling-32bit-binary-expected-unqualified-id-before-int128.

Open GNUMakefile and scroll down to around line 50. Find:

ifneq ($(GCC42_OR_LATER),0)
  ifeq ($(IS_DARWIN),1)
    CXXFLAGS += -arch x86_64 -arch i386
  else
    CXXFLAGS += -march=native
  endif
endif


And change to:

ifneq ($(GCC42_OR_LATER),0)
  ifeq ($(IS_DARWIN),1)
    CXXFLAGS += -arch x86_64
  else
    CXXFLAGS += -march=native
  endif
endif


Jeff

Tim Burgess

unread,
Dec 11, 2013, 5:01:33 AM12/11/13
to cryptop...@googlegroups.com

Hi Jeff,

 

Many thanks for your suggestion. The closest match I could find in my GNUmakefile happens at line 38 and was:

 

ifneq ($(GCC42_OR_LATER),0)

ifeq ($(UNAME),Darwin)

CXXFLAGS += -arch x86_64 -arch i386

else

CXXFLAGS += -march=native

endif

endif

 

Replacing this with your suggested code then performing a make clean before attempting a build yields:

 

macbook_pro:cryptopp tim$ make static test

c++ -DNDEBUG -g -O2 -march=native -DCRYPTOPP_DISABLE_ASM -pipe -c 3way.cpp

:2565:no such instruction: `vzeroupper'

:2675:no such instruction: `vzeroupper'

:2694:no such instruction: `vzeroupper'

:517:no such instruction: `vzeroupper'

:630:no such instruction: `vzeroupper'

:772:no such instruction: `vzeroupper'

:793:no such instruction: `vzeroupper'

:811:no such instruction: `vzeroupper'

:949:no such instruction: `vzeroupper'

:970:no such instruction: `vzeroupper'

:988:no such instruction: `vzeroupper'

:1372:no such instruction: `vzeroupper'

:1403:no such instruction: `vzeroupper'

:1845:no such instruction: `vzeroupper'

:1876:no such instruction: `vzeroupper'

:1927:no such instruction: `vzeroupper'

:1935:no such instruction: `vzeroupper'

:1952:no such instruction: `vzeroupper'

:1960:no such instruction: `vzeroupper'

:1977:no such instruction: `vzeroupper'

:2552:no such instruction: `vzeroupper'

:2570:no such instruction: `vzeroupper'

make: *** [3way.o] Error 1

 

I assume that this isn’t what you get on your system and I’m curious about the difference in the BNUmakefile. Do you have any more suggestions, please?

 

Best wishes.

Tim Burgess

Mobile Mouse

unread,
Dec 11, 2013, 7:07:07 AM12/11/13
to Tim Burgess, Crypto++ Mailing List
In addition to that change I suggest (assuming that you have one of the later Intel iCore processors) adding to CXXFLAGS the following:

-msse2 -mssse3 -msse4 -msse4.2 -maes -mpclmul



--
--
You received this message because you are subscribed to the "Crypto++ Users" Google Group.
To unsubscribe, send an email to cryptopp-user...@googlegroups.com.
More information about Crypto++ and this group is available at http://www.cryptopp.com.
---
You received this message because you are subscribed to the Google Groups "Crypto++ Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cryptopp-user...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

Tim Burgess

unread,
Dec 13, 2013, 5:20:33 AM12/13/13
to cryptop...@googlegroups.com
Hi,
 
I now have:
 
ifneq ($(GCC42_OR_LATER),0)
ifeq ($(IS_DARWIN),1) 
CXXFLAGS += -arch x86_64 -msse2 -mssse3 -msse4 -msse4.2 -maes -mpclmul
else
CXXFLAGS += -march=native
endif
endif
 
My Mac is about 2.5 years old and is a quad core Intel. Unfortunately, this change doesn't seem to change the build output - I'm still getting unknown instruction messages.


On Friday, 6 December 2013 16:21:49 UTC, Tim Burgess wrote:

David Irvine

unread,
Dec 13, 2013, 5:40:44 AM12/13/13
to Tim Burgess, Crypto++ Users

On Fri, Dec 13, 2013 at 10:20 AM, Tim Burgess <t...@raisedbar.net> wrote:
My Mac is about 2.5 years old and is a quad core Intel. Unfortunately, this change doesn't seem to change the build output - I'm still getting unknown instruction messages.

if you do 'c++ --version' is it definitely gcc 4.8. We build cryptop with clang on osx now and it's fine. I am not sure you can do the same but it may be a simpler mechanism if you intend to distribute code in osx later on. GCC 4.8 should also work though.

Best Regards
David Irvine

It's not difficult, it's different !

maidsafe.net Limited is a limited liability company incorporated in Scotland with number SC297540. VAT Registered 889 0608 77. Registered Office: 72 Templehill, Troon, KA10 6BE.
Telephone Scotland: +44 
 1292 317311

Tim Burgess

unread,
Dec 13, 2013, 6:10:43 AM12/13/13
to cryptop...@googlegroups.com
Hi,
 
Unfortunately I'm forced to use GCC as I'm using other libraries in my actual XCode 4.6 application project that still have some issues with clang. In fact my overall project seems to be plagued by version issues as I have one module that will only build using Apple's provided GCC 4.2 variant, as its written using Objective-C++ and GCC 4.8's Objective-C++ implementation doesn't support some of Apple's methods. The rest of my modules require GCC 4.8 as Apple's 4.2 hits a documented issue when using Boost shared pointers to an object containing variants. Anyway, that background isn't really relevant here. My C++ version output is:
 

macbook_pro:~ tim$ c++ --version

c++ (MacPorts gcc48 4.8.2_0) 4.8.2

Copyright (C) 2013 Free Software Foundation, Inc.

This is free software; see the source for copying conditions.  There is NO

warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

I'm more than prepared to believe that I'm doing something wrong, but don't have the experience on this platform to figure out the solution. I've had no issues with Crypto++ on the Windows build of my application.

 

Thanks again.
 
Tim
 

On Friday, 6 December 2013 16:21:49 UTC, Tim Burgess wrote:

David Irvine

unread,
Dec 13, 2013, 6:42:46 AM12/13/13
to Tim Burgess, Crypto++ Users

On Fri, Dec 13, 2013 at 11:10 AM, Tim Burgess <t...@raisedbar.net> wrote:
Unfortunately I'm forced to use GCC as I'm using other libraries in my actual XCode 4.6 application project that still have some issues with clang. In fact my overall project seems to be plagued by version issues as I have one module that will only build using Apple's provided GCC 4.2 variant, as its written using Objective-C++ and GCC 4.8's Objective-C++ implementation doesn't support some of Apple's methods. The rest of my modules require GCC 4.8 as Apple's 4.2 hits a documented issue when using Boost shared pointers to an object containing variants. Anyway, that background isn't really relevant here. My C++ version output is:

Those are nightmare scenarios for sure :-) If you do
"make VERBOSE=1"
and take a peek at what's output then we can see what flags the makefile is picking up. This will help, as a kludge you can force this with a hacked makefile to get you out of a hole, but best if we can fix the makefile for this particular scenario.

Tim Burgess

unread,
Dec 13, 2013, 7:52:50 AM12/13/13
to cryptop...@googlegroups.com
OK, here's the output:
 

macbook_pro:cryptopp tim$ make VERBOSE=1

c++ -DNDEBUG -g -O2 -march=native -DCRYPTOPP_DISABLE_ASM -pipe -c

 

I'm looking at the code in the make file:
 
ifneq ($(GCC42_OR_LATER),0) 
ifeq ($(IS_DARWIN),1) 
CXXFLAGS += -arch x86_64 -msse2 -mssse3 -msse4 -msse4.2 -maes -mpclmul
else
CXXFLAGS += -march=native
endif
endif

 

and surely the first line is in the wrong sense (not equal), or have I placed the edits in the wrong place?

 

Cheers.
 
Tim

 

 

 
 
 
 

On Friday, 6 December 2013 16:21:49 UTC, Tim Burgess wrote:

Mobile Mouse

unread,
Dec 13, 2013, 8:43:50 AM12/13/13
to Tim Burgess, cryptop...@googlegroups.com
I'm not that big an expert on make syntax. All I can say is: I've built crypto++ 5.6.2 on Mac OS X v10.9 under GCC-4.8 and 4.8.2 from macports, and for successful compilation & reasonable AES performance the CXXFLAGS must somehow include those flags I listed. I also added configurable .dylib vs .so, and -dynamiclib flag for the shared library. I can post my edited makefile...

Sent from my iPad
--

Tim Burgess

unread,
Dec 13, 2013, 10:25:54 AM12/13/13
to cryptop...@googlegroups.com
Hi,
 
I know next to nothing about make files, though I think it's going to be nmy next topic of study if I can find a good tutorial - I'm a blind programmer and Apple's VoiceOver speech output screen-reader is a pain with XCode, so the command-line beckons for future projects. I'd be very grateful if you'd post your file - it can't hurt, even if it doesn't do any good.
 
Cheers.
 
Tim
On Friday, 6 December 2013 16:21:49 UTC, Tim Burgess wrote:

Mouse

unread,
Dec 13, 2013, 1:37:50 PM12/13/13
to Tim Burgess, Crypto++ Users
1. My environment variable CXXFLAGS (and CFLAGS) is set to:
"-maes -mpclmul -msse2 -mssse3 -msse4 -msse4.2"

2. My GNUmakefile for cryptopp-5.6.2 is attached. Hope it would do some good to somebody. It works on Mac OS X 10.8.4 and 10.9.

3. Note that since I use Macports, /opt/local/bin/gcc is in my path prior to /usr/bin/gcc, so the build picks up gcc-4.8 from Macports rather than the older gcc from Xcode.

4. You still need to use assembler from Xcode, because both coming with either GC-4.8 and /usr/bin/as are unable to parse new instructions. I'm attaching the "as" script that should replace  /opt/local/bin/as (save the original 28K as executable somewhere). 

Hope it helps.


--
--
You received this message because you are subscribed to the "Crypto++ Users" Google Group.
To unsubscribe, send an email to cryptopp-user...@googlegroups.com.
More information about Crypto++ and this group is available at http://www.cryptopp.com.
---
You received this message because you are subscribed to the Google Groups "Crypto++ Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cryptopp-user...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.



--
Regards,
Mouse

as
GNUmakefile

Tim Burgess

unread,
Dec 13, 2013, 4:37:26 PM12/13/13
to cryptop...@googlegroups.com
OK, that gets me further, but not all of the way, here's the output:
 

macbook_pro:cryptopp tim$ make static test

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c 3way.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c adler32.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c algebra.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c algparam.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c arc4.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c asn.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c authenc.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c base32.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c base64.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c basecode.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c bfinit.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c blowfish.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c blumshub.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c camellia.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c cast.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c casts.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c cbcmac.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c ccm.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c channels.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c cmac.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c cpu.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c crc.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c cryptlib.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c cryptlib_bds.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c default.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c des.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c dessp.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c dh.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c dh2.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c dll.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c dsa.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c eax.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c ec2n.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c eccrypto.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c ecp.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c elgamal.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c emsa2.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c eprecomp.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c esign.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c files.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c filters.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c fips140.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c fipstest.cpp

g++ -DNDEBUG -g -O2 -ansi -fPIC -ffunction-sections -fdata-sections -maes -mpclmul -mtune=core-avx2 -arch x86_64  -DCRYPTOPP_DISABLE_ASM -c gcm.cpp

:50:no such instruction: `pclmulqdq $1, %xmm5,%xmm9'

:64:no such instruction: `pclmulqdq $1, %xmm4,%xmm8'

:69:no such instruction: `pclmulqdq $1, %xmm10,%xmm2'

:42:no such instruction: `pclmulqdq $16, %xmm5,%xmm7'

:48:no such instruction: `pclmulqdq $16, %xmm4,%xmm6'

:66:no such instruction: `pclmulqdq $16, %xmm10,%xmm3'

:52:no such instruction: `pclmulqdq $0, %xmm5,%xmm13'

:58:no such instruction: `pclmulqdq $1, %xmm4,%xmm1'

:89:no such instruction: `pclmulqdq $0, %xmm10,%xmm0'

:162:no such instruction: `pclmulqdq $16, %xmm14,%xmm2'

:195:no such instruction: `pclmulqdq $0, %xmm14,%xmm13'

:53:no such instruction: `pclmulqdq $16, %xmm5,%xmm13'

:59:no such instruction: `pclmulqdq $17, %xmm4,%xmm1'

:91:no such instruction: `pclmulqdq $16, %xmm10,%xmm0'

:63:no such instruction: `pclmulqdq $0, %xmm3,%xmm4'

:95:no such instruction: `pclmulqdq $16, %xmm3,%xmm2'

:102:no such instruction: `pclmulqdq $1, %xmm3,%xmm0'

:144:no such instruction: `pclmulqdq $16, %xmm5,%xmm6'

:169:no such instruction: `pclmulqdq $17, %xmm3,%xmm2'

:187:no such instruction: `pclmulqdq $0, %xmm5,%xmm4'

:298:no such instruction: `pclmulqdq $0, %xmm3,%xmm4'

:313:no such instruction: `pclmulqdq $16, %xmm3,%xmm1'

:319:no such instruction: `pclmulqdq $1, %xmm3,%xmm0'

:350:no such instruction: `pclmulqdq $17, %xmm3,%xmm1'

:367:no such instruction: `pclmulqdq $16, %xmm5,%xmm2'

:404:no such instruction: `pclmulqdq $0, %xmm5,%xmm2'

make: *** [gcm.o] Error 1

 

Many thanks for this help - I'll also take a look at the O'Reilly book.


On Friday, 6 December 2013 16:21:49 UTC, Tim Burgess wrote:

Mobile Mouse

unread,
Dec 13, 2013, 6:18:16 PM12/13/13
to Tim Burgess, Crypto++ Mailing List
To me it says that you haven’t installed assembler script that I posted. Or you installed it not on the path where gcc/g++ is looking for it.


Tim Burgess

unread,
Dec 17, 2013, 10:16:12 AM12/17/13
to cryptop...@googlegroups.com
OK, I'd made an error in my path. I corrected it and now all of the object (.o) files build without errors. Additionally, doing a make static test runs the test phase and says that all tests passed. However, there appear to be some problems during linkage, as I got several messages about modules containing no symbols, so I'm unsure as to whether I can trust my library .a file. The output is long, so I won't post it here, but I'd appreciate any feedback on the contents of:
 
 
Many thanks for the help so far - it's very much appreciated.
 
Tim

On Friday, 6 December 2013 16:21:49 UTC, Tim Burgess wrote:

Mouse

unread,
Dec 17, 2013, 11:29:42 AM12/17/13
to Tim Burgess, Crypto++ Users
I didn't even bother looking at that log. If you can successfully build and "cryptest.exe", it is proof enough that it works. And if you can build a few of your own sample applications, that's it. No need to worry any further.


--
--
You received this message because you are subscribed to the "Crypto++ Users" Google Group.
To unsubscribe, send an email to cryptopp-user...@googlegroups.com.
More information about Crypto++ and this group is available at http://www.cryptopp.com.
---
You received this message because you are subscribed to the Google Groups "Crypto++ Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cryptopp-user...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.



--
Regards,
Mouse

Tim Burgess

unread,
Dec 17, 2013, 12:10:59 PM12/17/13
to cryptop...@googlegroups.com
OK, that's reassuring. I'll incorporate the library into my application and report back if I hit any issues.
 
Again, many thanks for the help.
 
Cheers.
 
Tim

On Friday, 6 December 2013 16:21:49 UTC, Tim Burgess wrote:
Reply all
Reply to author
Forward
0 new messages