[llvm-dev] LLVM Weekly - #180, Jun 12th 2017

[Alex Bradbury via llvm-dev]

LLVM Weekly - #180, Jun 12th 2017
=================================

If you prefer, you can read a HTML version of this email at
<http://llvmweekly.org/issue/180>.

Welcome to the one hundred and eightieth issue of LLVM Weekly, a weekly
newsletter (published every Monday) covering developments in LLVM, Clang, and
related projects. LLVM Weekly is brought to you by [Alex
Bradbury](http://asbradbury.org). Subscribe to future issues at
<http://llvmweekly.org> and pass it on to anyone else you think may be
interested. Please send any tips or feedback to <asb at asbradbury.org>, or
@llvmweekly or @asbradbury on Twitter.


## News and articles from around the web

Rust verison 1.18 [has been
released](https://blog.rust-lang.org/2017/06/08/Rust-1.18.html). Rust is of
course one of a growing number of interesting programming languages that use
LLVM for code generation.


## On the mailing lists

* Duncan P. N. Exon Smith
[reports](http://lists.llvm.org/pipermail/cfe-dev/2017-June/054108.html) that
after the recently WWDC, Apple are intending to start the process of
contributing two new clang features: indexing-while-building and refactoring
support.

* Nuno Lopes
[proposes](http://lists.llvm.org/pipermail/llvm-dev/2017-June/113781.html)
changing undef to poison for a number of cases, e.g.
insertelement/extractelement with indices out of range.

* Peter Collingbourne has an [RFC on adding a one definition rule checker to
Clang and
LLD](http://lists.llvm.org/pipermail/llvm-dev/2017-June/113820.html).

* Stephen Crane has [proposed
pagerando](http://lists.llvm.org/pipermail/llvm-dev/2017-June/113794.html),
which randomises the location of pages containing code. This goes a step
further than address space layout randomisation (ASLR), which randomises just
the base address. The RFC goes in to quite some detail on how the mechanism
works, and you may also be interested in
[the paper](http://www.ics.uci.edu/~perl/sd16_pagerando.pdf).

* Clement Courbet has [requested
comments](http://lists.llvm.org/pipermail/llvm-dev/2017-June/113837.html) on a
new pass for optimising comparison chains.

* Charles Saternos has [shared an
RFC](http://lists.llvm.org/pipermail/llvm-dev/2017-June/113632.html) on the
format for the ThinLTO module summary in llvm-dis.

* Geoff Berry is
[hoping](http://lists.llvm.org/pipermail/llvm-dev/2017-June/113959.html) to
switch EarlyCSE function simplification to use MemorySSA by default. He is
looking for help testing this change.


## LLVM commits

* The AMDGPU backend gained a huge amount of new documentation.
[r304831](http://reviews.llvm.org/rL304831).

* Initial support for the DWARFv5 string offsets table was added.
[r304759](http://reviews.llvm.org/rL304759).

* The LLVM language reference is now explicit that shifting a value by an
amount greater than its bitwidth results in poison.
[r304780](http://reviews.llvm.org/rL304780).

* The vast majority of `#include` lines in files in the include/ and lib/ have
been sorted by filename. [r304787](http://reviews.llvm.org/rL304787).


## Clang commits

* `#pragma clang section` is now supported, and can be used to specify section
names for global variables. [r304705](http://reviews.llvm.org/rL304705).

* A new clang-tidy checker has been introduced which will replace dynamic
exception specifications with `noexcept` or a user-defined macro.
[r304977](http://reviews.llvm.org/rL304977).


## Other project commits

* A new test framework has been added for testing LLDB's client-server
communication. [r304793](http://reviews.llvm.org/rL304793).

* LLVM's OpenMP runtime gained support for OpenMP 4.5's
`schedule(simd:guided)` and `schedule(simd:runtime)`.
[r304724](http://reviews.llvm.org/rL304724).