Re: [Boost-bugs] [Boost C++ Libraries] #9766: boost >= 1.54 failes to compile with gcc-4.8.2 and LTO enabled

Subject: Re: [Boost-bugs] [Boost C++ Libraries] #9766: boost >= 1.54 failes to compile with gcc-4.8.2 and LTO enabled
From: Boost C++ Libraries (noreply_at_[hidden])
Date: 2016-02-14 11:10:59


#9766: boost >= 1.54 failes to compile with gcc-4.8.2 and LTO enabled
-------------------------------+---------------------
  Reporter: steffen@… | Owner: andysem
      Type: Bugs | Status: closed
 Milestone: To Be Determined | Component: log
   Version: Boost 1.54.0 | Severity: Problem
Resolution: fixed | Keywords:
-------------------------------+---------------------

Comment (by anonymous):

 Actually, I'm using gcc-5.3, and it still causes build failure on amdfam10
 (AMD Phenom II). I'm arguing that overriding -march isn't one of the
 compiler options that should be set for different translation units. This
 is something that often crops up in Gentoo, where -march is set system-
 wide, and then gets overridden by upstream build-systems not following
 best practises. There are plenty of patches in Portage to work around
 these issues.

 The simplest solution is to just make the optimized functions individually
 build-time configurable, that way a generic distributor build can chose a
 suitable generic -march, along with enabling run-time CPU optimizations.
 Conversely, a targeted build can just include support and optimization as
 appropriate.

 I brought up the function attributes because that's the solution GCC has
 implemented to support this type of situation as was discussed in the GCC
 bug report, but as you say, it's only good for GCC>4.9.

-- 
Ticket URL: <https://svn.boost.org/trac/boost/ticket/9766#comment:21>
Boost C++ Libraries <http://www.boost.org/>
Boost provides free peer-reviewed portable C++ source libraries.

This archive was generated by hypermail 2.1.7 : 2017-02-16 18:50:19 UTC