[llvm-dev] LLVM Weekly - #314, Jan 6th 2020

[Alex Bradbury via llvm-dev]

LLVM Weekly - #314, Jan 6th 2020
================================

If you prefer, you can read a HTML version of this email at
<http://llvmweekly.org/issue/314>.

Welcome to the three hundred and fourteenth 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](https://www.linkedin.com/in/alex-bradbury/). 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.

The very [first LLVM Weekly issue](http://llvmweekly.org/issue/1) went out six
years ago today, on Jan 6th 2014. Since then there's been a new issue every
single Monday, with zero gaps. Thank you everyone for reading so far! I of
course intend to keep it going through 2020 as well.


## News and articles from around the web

The deadline for proposals for EuroLLVM 2020 [is this Saturday 11th
January](http://lists.llvm.org/pipermail/llvm-dev/2020-January/137881.html).
EuroLLVM will take place April 6th-7th in Paris.

The program for the LLVM dev room at FOSDEM 2020 [has been
published](https://fosdem.org/2020/schedule/track/llvm/).

Corentin Jabot wrote up a detailed blog post on [improving performance for
large integer arrays in Clang](https://cor3ntin.github.io/posts/arrays/).

The second LLVM social in Bangalore [will take
place](http://lists.llvm.org/pipermail/llvm-dev/2020-January/137938.html) on
Saturday February 1st.


## On the mailing lists

* Mark de Wever kicked off an RFC thread on [handling implementation limits in
Clang](http://lists.llvm.org/pipermail/cfe-dev/2020-January/064177.html),
suggesting that implementation limits of the Clang compiler be centralised to
allow them to be representated in generated headers and documentation.

* Michael Kruse [proposes a new loop optimisation
framework](http://lists.llvm.org/pipermail/llvm-dev/2020-January/137909.html).
"The central idea is to use a modifiable loop tree - similar to LoopInfo - as
the primary representation. LLVM-IR is converted to a loop tree, then
optimized and finally LLVM-IR is generated again for subtrees that are
considered profitable."

* David Blaikie posted an RFC on [reducing .o file size in
DWARFv5](http://lists.llvm.org/pipermail/llvm-dev/2019-December/137863.html),
Using `DW_AT_ranges` can resuled in significant size savings for for .o, but
can cause a very small increase for .dwo files.

* Whitney T Tsang proposes [changing LoopUnrollAndJamPass from a loop to a
function
pass](http://lists.llvm.org/pipermail/llvm-dev/2020-January/137901.html).


## LLVM commits

* The pass registration mechanism used by Polly was generalised so it can be
used by any third party tool.
[24ab9b5](https://reviews.llvm.org/rG24ab9b537e6).

* The check for the `disable-tail-calls` attribute was hoisted from individual
backends to the instruction selectors.
[9c2b728](https://reviews.llvm.org/rG9c2b72821be).

* The NoFPExcept SDNodeFlag now defaults to false.
[6333679](https://reviews.llvm.org/rG63336795f0d).

* Zlib support is now disabled by default on Windows.
[a2ca1c2](https://reviews.llvm.org/rGa2ca1c2d566).


## Clang commits

* OpenMP 5.0 implementation continues with support for codegen of lastprivate
conditional list items. [a58da1a](https://reviews.llvm.org/rGa58da1a2ff0).

* clang-tidy gained a utility function to add 'const' to variables.
[cf48101](https://reviews.llvm.org/rGcf48101200e).


## Other project commits

* MLIR gained documentation for the VectorOps dialect.
[a932f03](https://reviews.llvm.org/rGa932f033a34).

* LLD gained support for TPREL relocations for Hexagon.
[81ffe89](https://reviews.llvm.org/rG81ffe89735e).