[LLVMdev] LLVM Weekly - #66, Apr 6th 2015

[Alex Bradbury]

LLVM Weekly - #66, Apr 6th 2015
===============================

If you prefer, you can read a HTML version of this email at
<http://llvmweekly.org/issue/66>.

Welcome to the sixty-sixth 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

`color_coded`, a vim plugin for syntax highlighting using libclang is
[now available](https://github.com/jeaye/color_coded).

Ravi, a dialect of Lua with JIT compilation via LLVM has [has its first alpha
release](https://github.com/dibyendumajumdar/ravi/releases). The status of JIT
compilation can be seen
[here](http://the-ravi-programming-language.readthedocs.org/en/latest/ravi-jit-status.html).


## On the mailing lists

* James Knight is [asking for advice on supporting 64-bit load/store on a
32-bit arch](http://article.gmane.org/gmane.comp.compilers.llvm.devel/84258).
Respondents point to examples from ARM and R600.

* Eric Christopher has [kicked off another discussion on LTO and codegen
options](http://article.gmane.org/gmane.comp.compilers.llvm.devel/84144).

* Katya Romanova [proposes adding doxygen comments for
intrinsics](http://article.gmane.org/gmane.comp.compilers.clang.devel/41911).
The intention is that the current documentation would be converted in an
automated way. So far, people seem to be in favour.

* Douglas Gregor is [stepping down as code owner of 'all parts of Clang not
covered by someone
else'](http://article.gmane.org/gmane.comp.compilers.clang.devel/42046).
Richard Smith will be taking over. Thank you Douglas for the years of hard
work.

* Can you [cross-compile LLVM's test
suite?](http://article.gmane.org/gmane.comp.compilers.llvm.devel/84251). The
answer, is yes.

* Duncan P.N. Exon Smith proposes that
[preserve-bc-use-list-order](http://article.gmane.org/gmane.comp.compilers.llvm.devel/84196)
be on by default.


## LLVM commits

* API migration has started for GEP constant factories. For now, nullptr can
be passed for the pointee type, but you'll need to pass the type explicitly to
be future-proof. [r233938](http://reviews.llvm.org/rL233938).

* A proof of concept fuzzer based on DataFlowSanitizer has been added, as well
as support for token-based fuzzing.
[r233613](http://reviews.llvm.org/rL233613),
[r233745](http://reviews.llvm.org/rL233745).

* DebugLoc's API has been rewritten.
[r233573](http://reviews.llvm.org/rL233573).

* The SystemZ backend now supports transactional execution on the zEC12.
[r233803](http://reviews.llvm.org/rL233803).


## Clang commits

* Clang gained a toolchain driver for targeting NaCl.
[r233594](http://reviews.llvm.org/rL233594).

* The size of various Stmt subclasses has been optimised on 64-bit targets.
[r233921](http://reviews.llvm.org/rL233921).

* Codegen was added for the OpenMP atomic update construct.
[r233513](http://reviews.llvm.org/rL233513).


## Other project commits

* LLDB system initialization has been reworked.
[r233758](http://reviews.llvm.org/rL233758).