<div dir="ltr"><br><div class="gmail_extra"><br><div class="gmail_quote">On Mon, Apr 27, 2015 at 5:51 PM, Geof Sawaya <span dir="ltr"><<a href="mailto:fredericflintstone@gmail.com" target="_blank">fredericflintstone@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr">David, thank you for responding to my question<br><div class="gmail_extra"><br><div class="gmail_quote"><span class="">On Thu, Apr 16, 2015 at 6:09 PM, David Blaikie <span dir="ltr"><<a href="mailto:dblaikie@gmail.com" target="_blank">dblaikie@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:1ex">It's probably not VM related, but non-asserts builds of LLVM don't<br>
produce named values. It's only the +Asserts builds of LLVM that<br>
produce value (& block, I guess) names.<br></blockquote><div><br></div></span><div> Unfortunately, I do have asserts enabled in my build (LLVM 3.2) and am getting this described name stripping.</div><div><br></div><div>Any other ideas shared about avoiding this problem will be greatly appreciated.</div></div></div></div></blockquote><div><br>Nothing else rings a bell - a small example might be helpful, if you have one.<br> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr"><div class="gmail_extra"><div class="gmail_quote"><span class=""><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:1ex">
<br>
This is an optimization to avoid the extra work when we're just doing<br>
the actual work of compiling anyway.<br>
<br>
It'd be best to build your tool to not rely on these names.<br></blockquote><div><br></div></span><div>Yes, this should be done in the future, I imagine. I guess that the information I need to preserve will have to be stored in metadata? Not worried about this now as long as I can keep the label / block names for now.</div></div></div></div></blockquote><div><br>*nod*<br> </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr"><div class="gmail_extra"><div class="gmail_quote"><div><br></div><div>Thanks -- Geof</div><span class=""><div><br></div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:1ex">
<div><div><br>
On Thu, Apr 16, 2015 at 1:43 PM, Geof Sawaya<br>
<<a href="mailto:fredericflintstone@gmail.com" target="_blank">fredericflintstone@gmail.com</a>> wrote:<br>
> Hi Devs,<br>
><br>
> I'm developing a tool that relies on semantic information in bytecode labels<br>
> (i.e. block names).<br>
><br>
> I've discovered that clang is stripping these named labels (along with some<br>
> virtual register names) when I run on a virtual machine. Well, I'm using<br>
> VirtualBox, and have tried two different versions of Ubuntu and some<br>
> different clang builds.<br>
><br>
> Can someone point me in the right direction to understand why the IR would<br>
> be emitted differently because clang is running on a VM?<br>
><br>
> Many thanks -- Geof<br>
><br>
</div></div>> _______________________________________________<br>
> LLVM Developers mailing list<br>
> <a href="mailto:LLVMdev@cs.uiuc.edu" target="_blank">LLVMdev@cs.uiuc.edu</a> <a href="http://llvm.cs.uiuc.edu" target="_blank">http://llvm.cs.uiuc.edu</a><br>
> <a href="http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev" target="_blank">http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev</a><br>
><br>
</blockquote></span></div><br></div></div>
</blockquote></div><br></div></div>