<html><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; "><br><div><div>On Dec 19, 2007, at 2:25 PM, Anton Korobeynikov wrote:</div><br><blockquote type="cite">Chris,<br><br><blockquote type="cite">Nice! What is the magic needed to trick it into doing this? I'd <br></blockquote><blockquote type="cite">like the nightly testers to do this. If you tell me the majik, I can <br></blockquote><blockquote type="cite">try to hook it up to happen automatically.<br></blockquote>Usual magic is to generate .S file and compile/link it with native g++<br></blockquote></div><div><br class="webkit-block-placeholder"></div><div>(or .o file)</div><div>That works on an individual file basis. The problem file is libgcc_s.10.5.dylib</div><div>(would be 10.4 on Tiger, though I haven't exercised that). Removing that from</div><div>llvm-gcc.obj/gcc and anywhere in the ld search page you might have installed it,</div><div>such as /usr/local/lib, is a necessary step (the testsuite harness passes in its</div><div>own -L and I haven't found a way to override it). There is a good dylib in /usr/lib.</div><div><br class="webkit-block-placeholder"></div><div>However, there's more. To take a concrete example, I'll discuss g++.dg/eh/ctor1.C. </div><div>In this case, the bad unwinding code will hang at runtime, while compiling from the</div><div>command line with the local llvm-gcc and a good unwinding library works correctly.</div><div><br class="webkit-block-placeholder"></div><div>However, I have not been able to get things to work using 'make check RUNTESTFLAGS=...'</div><div>even though cutting and pasting the failing compilation line from the log file works fine,</div><div>and even though passing in -Wl,-t shows the same libraries are being found in both cases. </div><div>The failing behavior is different: it says</div><div><br class="webkit-block-placeholder"></div><div><b>terminate called after throwing an instance of 'int'</b></div><div><b>terminate called recursively</b></div><div><b><br class="webkit-block-placeholder"></b></div><div>rather than looping. This looks more like EH is not being turned on somehow. </div><div>For your comfort and convenience, the harness deletes all the interesting files before</div><div>exiting, so tracking this down is daunting. I'm sure this is a question of invoking the</div><div>testsuite properly; can any testsuite gurus help?</div><div><br></div></body></html>