On Thursday, February 12, 2015, Sergey Dmitrouk <<a href="mailto:sdmitrouk@accesssoftek.com">sdmitrouk@accesssoftek.com</a>> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">On Wed, Feb 11, 2015 at 09:46:36PM -0800, Saleem Abdulrasool wrote:<br>
> Can you provide more information about how you are building things?A<br>
> Perhaps an exact set of steps to reproduce this.A The change should only<br>
> effects unwind.A As such, the following is sufficient for checking the<br>
> result:<br>
> $ cat reduced.cc<br>
> extern "C" int printf(const char *, ...);<br>
> int main(int argc, char **argv) {<br>
> A try { throw 1; } catch (const int &i) { printf("caught %u\n", i); }<br>
> A return 0;<br>
> }<br>
> $ armv7a-hardfloat-linux-gnueabi-g++-4.8.3 -fno-stack-protector -c<br>
> reduced.cc -o reduced.o<br>
> $ armv7a-hardfloat-linux-gnueabi-g++-4.8.3 -nodefaultlibs reduced.o -o<br>
> reduced -Llib -lunwind -lc -lc++abi<br>
> $ LD_LIBRARY_PATH=lib ./reduced<br>
> caught 1<br>
> That was with current ToT of libc++abi and libunwind (LLVM).<br>
<br>
Sorry for bothering, it doesn't happen if I compile libunwind with gcc. Only<br>
libunwind built with clang fails, so it must be an issue somewhere else.</blockquote><div><br></div>I've been successful with clang as well.<div><br></div><div>Reproduction steps would help in identifying what is happening.<span></span></div><div><div> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Thanks,<br>
Sergey<br>
</blockquote></div><br><br>-- <br>Saleem Abdulrasool<br>compnerd (at) compnerd (dot) org<br>