[llvm-dev] LLVM Weekly - #311, December 16th 2019

Alex Bradbury via llvm-dev llvm-dev at lists.llvm.org
Mon Dec 16 12:48:57 PST 2019

LLVM Weekly - #311, December 16th 2019

If you prefer, you can read a HTML version of this email at

Welcome to the three hundred and eleventh 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.

## News and articles from around the web

All videos from the 2019 LLVM Developers' Meeting are now posted and available
in a [handy YouTube

LLVM 9.0.1-rc3 [has been

## On the mailing lists

* Teresa Johnson shared an RFC on [supporting agressive whole program
devirtualization without

* Jonas Devlieghere wrote up an RFC on [supporting Lua scripting in

* Sam McCall starts a [discussion on providing network RPC in LLVM
and is seeking feedback on depending on external libraries like Thrift, grpc

* Sander De Smalen [shares notes from last week's SVE sync-up
call](http://lists.llvm.org/pipermail/llvm-dev/2019-December/137615.html) and
the agenda for the next one.

* Raphael Isemann shared a summary detailing [the future of modern-type-lookup
in LLDB](http://lists.llvm.org/pipermail/lldb-dev/2019-December/015831.html).
"modern-type-lookup is a flag that makes LLDB use `clang::ExternalASTMerger`
instead of directly using the `clang::ASTImporter` via our ClangASTImporter

## LLVM commits

* Initial matrix intrinsics and a lowering pass were added.

* Initial work related to profile guided size opimization was committed. This
adds instrumentation, but doesn't yet enable size optimisation.

* Support for a Vector Function Database was merged.

* The Arm cost model now understands cases where a shift can be folded into
other instructions. [be7a107](https://reviews.llvm.org/rGbe7a1070700).

* New tests were added for resolving of RISC-V relocations.

* Dead loop update instructions are now removed from Arm low overhead loops.

* getIntImmCost was renamed to getIntImmCostInst and getIntImmCostIntrin.

* Target-specific intrinsic enums were split into separate headers. This saves
222MB of object file size in a Windows optimised+debug build.

## Clang commits

* A new TableGen backend was added for generating serialisation classes.

* Clang support was added for constrained FP builtins.

* Clang now defaults to `-fuse-init-array`, using `-fno-use-init-array` only
for platforms known not to support `.init_array`.

* Clang learned to synthesise defaulted comparison functions for C++20.

## Other project commits

* A new script was added for symbolising hwasan reports after the fact.

* Fuzzing was added for parts of `<random>` in libcxx.

More information about the llvm-dev mailing list