[pkg-boost-commits] r14484 - boost/trunk/debian
Steven Michael Robbins
smr at alioth.debian.org
Sat Sep 5 17:50:13 UTC 2009
Author: smr
Date: 2009-09-05 17:50:13 +0000 (Sat, 05 Sep 2009)
New Revision: 14484
Modified:
boost/trunk/debian/Notes
Log:
Remove outdated info
Modified: boost/trunk/debian/Notes
===================================================================
--- boost/trunk/debian/Notes 2009-09-05 17:46:16 UTC (rev 14483)
+++ boost/trunk/debian/Notes 2009-09-05 17:50:13 UTC (rev 14484)
@@ -1,40 +1,11 @@
- GCC 4.3 Patches
- ---------------
-
-As of 2009-02-21, there are a number of issues with respect to GCC
-4.3. A post on the Boost list gives a query to list all the open
-reports in the Boost trac system:
-
- http://svn.boost.org/trac/boost/query?status=new&status=assigned&status=reopened&summary=~4.3&order=priority
-
-Here we track the open tickets and status with respect to the Debian
-patch series.
-
- Ticket Inc? Comment
- ------ ---- ----------------------------------
- 1605 no Library property_tree not present in 1.38
- 1955 no Library property_tree not present in 1.38
- 2321 no Warning
- 2388 no Warning
- 2390 no Warning
- 2582 yes boost_python_translate_exception.patch
- 2680 no Test suite failure, no patch.
-
-
Thread Support
--------------
All debian libraries are supposed to be built with _REENTRANT defined
-(policy 10.2). For the jam-built libraries, use "<define>_REENTRANT"
-on command line in debian/rules. For graph library, set in Makefile.
+(policy 10.2). We thus use "<define>_REENTRANT" on the bjam command
+line in debian/rules.
-The __sync functions are giving problems. I think that atomic_count.hpp
-is finally patched. It looks like sp_counted_base.hpp is still a
-problem on m68k. Revisit the strategy of using DBOOST_SP_USE_PTHREADS
-for the arches that don't have __sync; upsteam now has a spinlock-based
-default version of sp_counted_base.
-
Documentation
-------------
More information about the pkg-boost-commits
mailing list