Compiling ACE+TAO on Mac OS X

So, what's ACE+TAO?

It's big, it's complex, it's middleware!


In trying to compile ACE 5.7.5 and TAO 1.7.5 on my Mac OS X 10.5.8 system I encountered some strange error messages while linking. These were duplicate symbol errors I couldn't really track down to their root cause but was able to find simple cures.

For those poor souls also hitting those faults the following list of changes may help:
  1. File $ACE_ROOT/ace/Based_Pointer_Repository.h
  2. File $ACE_ROOT/ace/Obstack.h
Re 1. : File $ACE_ROOT/ace/Based_Pointer_Repository.h

I had to comment out the ACE_SINGLE_DECLARATION on lines 82 to 84:

/* - frgo, 2010-01-03
/// Declare a process wide singleton
ACE_SINGLETON_DECLARE (ACE_Singleton,
ACE_Based_Pointer_Repository,
ACE_SYNCH_RW_MUTEX)
*/

Re 2. : File $ACE_ROOT/ace/Obstack.h

I had to comment out the ACE_SINGLE_DECLARATION on line 24:

// - frgo, 2010-01-03
// ACE_SINGLETON_DECLARATION (ACE_Obstack_T ;)

Having solved these problems I was able to build and link ACE. Now on to TAO...

TAO also exhibits duplicate symbol errors. To get rid of these the following files have to be edited:
  1. File $TAO_ROOT/orbsvcs/orbsvcs/AV/AV_Core.h
  2. File $TAO_ROOT/orbsvcs/tests/Notify/lib/LookupManager.h
Re 1. File $TAO_ROOT/orbsvcs/orbsvcs/AV/AV_Core.h

I had to comment out TAO_AV_SINGLE_DECLARE on line 167:

// - frgo, 2010-01-03
// TAO_AV_SINGLETON_DECLARE (ACE_Singleton, TAO_AV_Core, ACE_Null_Mutex)

Re 2. File $TAO_ROOT/orbsvcs/tests/Notify/lib/LookupManager.h

I had to comment out TAO_NOTIFY_TEST_SINGLETON_DECLARE on line 113:

// - frgo, 2010-01-03// - frgo, 2010-01-03
// TAO_NOTIFY_TEST_SINGLETON_DECLARE (ACE_Singleton, TAO_Notify_Tests_LookupManager, TAO_SYNCH_MUTEX)

Having implemented these changes you can run

$ make install

and you are up and running with the TAO ORB.

Cheers

Frank

Kommentare

Unknown hat gesagt…
Commenting out these could lead to runtime problems.
3. Januar 2010 um 23:43
frgo hat gesagt…
Yep. Very true. So far no problems, though. If there's a better solution I'd be very interested to hear!

Thanks for pointing this out!

Regards
Frank
11. Januar 2010 um 05:24
SWS hat gesagt…
I've run into a similar problem, and I think it boils down to the compiler version. See
http://developer.apple.com/library/mac/#documentation/DeveloperTools/gcc-4.2.1/gcc/Visibility-Pragmas.html
and
http://developer.apple.com/library/mac/#documentation/DeveloperTools/Conceptual/CppRuntimeEnv/Articles/SymbolVisibility.html

In particular: “Although template declarations can be marked with the visibility attribute, template instantiations cannot. This is a known limitation and may be fixed in a future version of GCC.”
18. April 2011 um 13:41
Kommentar veröffentlichen

Beliebte Posts aus diesem Blog

SLIME, Lisp, Editing - the lost slime-close-parens-at-point

Once upon a time there was a rich SLIME version that had all bateries included in a few files. Now this old, old version is long gone. The Powers Of SLIME deleted the function slime-close-parens-at-point whose use and benefit is described in this blog post of Bill Clementson . For those requiring this exact feature here's the original code of that function: (setq slime-close-parens-limit 16) (defun slime-close-parens-at-point () "Close parenthesis at point to complete the top-level-form. Simply inserts ')' characters at point until `beginning-of-defun' and `end-of-defun' execute without errors, or `slime-close-parens-limit' is exceeded." (interactive) (loop for i from 1 to slime-close-parens-limit until (save-excursion (slime-beginning-of-defun) (ignore-errors (slime-end-of-defun) t)) do (insert ")"))) In addition to putting this in my .emacs file I have also the following line in it: (define-key slime-mode-map (kbd "...
Mehr anzeigen

The ADT-200A SDR TRX in practice ... A talk @ DARC P31

Bild
Dr YLs es OMs ! On March 23rd I gave a presentation of the features of the ADT-200A Software Defined Transceiver at the local Amateur Radio Club, DARC e.V. OV Ermstal DOK P31 .  The presentation focused on practical aspects of actually using the ADT-200A. We had about 15 radio amateurs (only OMs, unfortunately, HI) in the room and what was planned as a 1 hour talk turned into a 2 hour session ... We had fun using the TRX  and two CW QSOs have been completed with RA9XU - about 3000 km with just 50 watts into a dipole on 40 meters.  The exceptional performance of this TRX created some "wow" and "ah" effects -  it was a pleasure listening to CW signals with just 150 Hz bandwidth - the digital CW filters are going down to 50 Hz (we used 150 Hz for the QSOs) and there's no signal degradation whatsoever. Really sweet  ... One OM asked if I would by another TRX if I had all the money I can imagine - I just answered "Why should I ?" ... Fo...
Mehr anzeigen