[llvm-dev] LLVM Weekly - #198, Oct 16th 2017

[Alex Bradbury via llvm-dev]

LLVM Weekly - #198, Oct 16th 2017
=================================

If you prefer, you can read a HTML version of this email at
<http://llvmweekly.org/issue/198>.

Welcome to the one hundred and ninety-eighth 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.

Like many of you, I'll be attend the LLVM Dev Meeting this week - be sure to
say hi!


## News and articles from around the web

The [call for
speakers](http://lists.llvm.org/pipermail/llvm-dev/2017-October/118062.html)
has been issued for the LLVM Performance Workshop at CGO 2018. The deadline
for submissions is December 1st, 2017.

The [call for papers /
participation](http://lists.llvm.org/pipermail/llvm-dev/2017-October/118157.html)
has been posted for the LLVM toolchain devroom at FOSDEM 2018. This will take
place on Feb 4th 2018 in Brussels. The proposal deadline is Nov 26th, 2017.

The [CppCheck GUI](http://cppcheck.sourceforge.net/) has been
[extended](http://lists.llvm.org/pipermail/cfe-dev/2017-October/055709.html)
with support for clang-tidy. The author hopes to add support for the Clang
static analyzer in the future too. See
[here](http://cppcheck.sourceforge.net/clangtidy/index.html) for some
screenshots.


## On the mailing lists

* Lang Hames kicked off a discussion about improving [MC's memory ownership
conventions](http://lists.llvm.org/pipermail/llvm-dev/2017-October/118135.html).

* David Jones has posted an RFC on [adding a configuration option to enable
constexpr constructors in
C++1x/y](http://lists.llvm.org/pipermail/cfe-dev/2017-October/055724.html).

* Rui Ueyama is seeking input on an [issue with weak undefined symbols and
dynamic
libraries](http://lists.llvm.org/pipermail/llvm-dev/2017-October/118177.html).
Peter Smith started to [sketch
out](http://lists.llvm.org/pipermail/llvm-dev/2017-October/118237.html) a
solution.

* Kostya Serebryany has
[explained](http://lists.llvm.org/pipermail/llvm-dev/2017-October/118132.html)
the current policy on the use of the C++ standard library in the fuzzers.


## LLVM commits

* LLVM now has much more detailed documentation about the various fuzzers
available for fuzzing LLVM itself.
[r315544](http://reviews.llvm.org/rL315544).

* .Lcfi labels are no longer emitted when producing textual assembly.
[r315259](http://reviews.llvm.org/rL315259).

* The ARM backend now gives better diagnostics for floating point and general
purpose register operands. [r315303](http://reviews.llvm.org/rL315303),
[r315304](http://reviews.llvm.org/rL315304).

* The NumFixedArgs field of CallLoweringInfo is now correctly updated. As
noted in the commit message, there is a small chance this could affect your
backend if you were directly accessing this field or the IsFixed property.
[r315457](http://reviews.llvm.org/rL315457).

* MergeFunctions has been improved.
[r315852](http://reviews.llvm.org/rL315852),
[r315853](http://reviews.llvm.org/rL315853).

* createElfObjectWriter now takes a `std::unique_ptr<MCElfObjectTargetWriter>`
and returns a `std::unique_ptr<MCObjectWriter>`. If you have an out-of-tree
backend, it will need updating. The MCObjectStreamer prototype has also
changed. [r315254](http://reviews.llvm.org/rL315254),
[r315327](http://reviews.llvm.org/rL315327),
[r315531](http://reviews.llvm.org/rL315531).

* The loop invariant code motion pass will no longer make the illegal
optimisation of sinking an unordered atomic load into a loop.
[r315438](http://reviews.llvm.org/rL315438).


## Clang commits

* Clang gained support for destroying operator delete, as proposed for C+++2a.
[r315662](http://reviews.llvm.org/rL315662).

* The `-Wtautological-constant-compare` warning has been reintroduced.
[r315875](http://reviews.llvm.org/rL315875).


## Other project commits

* LLD has seen a large number of code cleanups. e.g.
[r315525](http://reviews.llvm.org/rL315525),
[r315617](http://reviews.llvm.org/rL31617).

* LLDB learnt to search for split DWARF files in a specified search path.
[r315387](http://reviews.llvm.org/rL315387).


## Review Corner

The LLVM Weekly review corner serves to highlight patches that are stuck
waiting awaiting review, or work from first-time contributors. See
[here](http://llvmweekly.org/reviewcorner) for more information and how to
submit you work for inclusion. Of course the hope is that highlighting these
patches will enable LLVM Weekly readers will step up and help to get them
merged. I'll be reporting back each week on any activity generated on these
patches, as well as sharing a new batch. If you want your patch included you
must submit it via the linked form.

Last week featured three patches, with two out of three seeing fresh review
activity after inclusion. [D36471](https://reviews.llvm.org/D36471) sadly has
seen no additional feedback. Thanks to Jessica Paquette and Anna Zaks for
contributing review comments to last week's patches.

There is just one patch submission this week:

* "Trigger a -Wtautological-pointer-compare warning on variables within
parentheses". [D38718](https://reviews.llvm.org/D38718), patch by 'erikv'.