|
From: | lilypond |
Subject: | Re: Issue 2013 in lilypond: clang error flower/file-name.cc |
Date: | Fri, 04 Nov 2011 14:39:10 +0000 |
Comment #8 on issue 2013 by address@hidden: clang error flower/file-name.cc http://code.google.com/p/lilypond/issues/detail?id=2013Please take a look at Issue 1875. The Boost fragment we pulled in and that is apparently used for just "make check" and nothing else, does an #include <cxxabi.h> in order to do C++ demangling for some reason.
That's a _vast_ amount of complexity we are pulling in just for doing a basic test. And it _fails_ with g++-4.6.1. Just the above-mentioned include makes it fail when linking, except when you omit to generate inline functions.
Is this supposed to be a regtest for Boost, for g++, or for Lilypond/flower? And what good is a regtest when nobody understands the complexity of the system and can interpret the failure correctly?
[Prev in Thread] | Current Thread | [Next in Thread] |