<div dir="ltr"><div class="gmail_extra"><br><div class="gmail_quote">On Tue, Nov 12, 2013 at 10:20 AM, <span dir="ltr"><<a href="mailto:dag@cray.com" target="_blank" class="cremed">dag@cray.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div class="im">Chris Lattner <<a href="mailto:clattner@apple.com" class="cremed">clattner@apple.com</a>> writes:<br>
<br>
>> If we say we support 4.7.x, then I don't think we should use c++11<br>
>> features that aren't supported and working on all 4.7.x versions.<br>
><br>
> Why not just explicitly say 4.7.1 and later? Are there any buildbots<br>
> that build with 4.7.1 yet? If not, that should be a prerequisite to<br>
> making the move.<br>
<br>
</div>Does that statement include 4.8.x and beyond?<br>
<br>
My main concern, as outlined in another message I just sent, is what<br>
"4.7.1 and later" means in terms of support from upstream. Do the<br>
inevitable problems discovered get worked around in the LLVM project<br>
sources or are problematic compilers simply noted as such in the<br>
documentation, potentially forcing users to upgrade their toolchain<br>
again?<br>
<br>
I can see tradeoffs with each strategy. It's important to be clear so<br>
we have proper expectations.<br></blockquote><div><br></div><div>I don't think we can predict this.</div><div><br></div><div>I think we clearly *want* to support all of the point releases that we can, but I think we would have to look at the *specific* work arounds for any toolchain bug discovered, as well as how widespread that toolchain was in terms of deployment. I can imagine both cases where we would easily work around the problem, and other cases where it would be completely unmaintainable.</div>
<div><br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">I'm pushing on this because although most toolchain upgrades are simple,<br>
when they go bad they go REALLY bad. It's a risk that has to be<br>
managed.</blockquote></div><br>While I understand that, the problem you're highlighting isn't new and isn't related to C++11. Discussing it here isn't really productive.</div></div>