[llvm-dev] LLVM Weekly - #330, Apr 27th 2020

[Alex Bradbury via llvm-dev]

LLVM Weekly - #330, Apr 27th 2020
=================================

If you prefer, you can read a HTML version of this email at
<http://llvmweekly.org/issue/330>.

Welcome to the three hundred and thirtieth 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

Erich Keane wrote for the LLVM blog about [the recently added `_ExtInt`
feature](http://blog.llvm.org/2020/04/the-new-clang-extint-feature-provides.html),
providing support for exact bitwidth integer types in Clang.


## On the mailing lists

* Tom Stellard returned to the topic of [switching from Bugzilla to GitHub
issues](http://lists.llvm.org/pipermail/llvm-dev/2020-April/141096.html),
reporting that support for subscribing to notifications for specific issue
labels will be provided through GitHub Actions. This spawned a long thread,
mostly focused on the question of how to handle existing bugs and
possibilities for keeping consistent numbering or using redirects.

* Martin Liška [announces
C-Vise](http://lists.llvm.org/pipermail/llvm-dev/2020-April/141188.html), a
Python fork of the C-Reduce tool.

* Kees Cook proposed [making -ftrivial-auto-var-init=zero a first-class
option](http://lists.llvm.org/pipermail/cfe-dev/2020-April/065221.html),
reporting a number of vendors using zero initialisation in production code
(with more responding later in the thread). There were many responses on this
thread and a consensus has yet to form, though JF Bastien indicated [interest
in seeing the work on optimising pattern initialisation
finished](http://lists.llvm.org/pipermail/cfe-dev/2020-April/065242.html) to
allow a fuller comparison.

* Sjoerd Meijer kicked off a [long thread on the clang mailing list about
mtune](http://lists.llvm.org/pipermail/cfe-dev/2020-April/065169.html). David
Greene [suggested not following GCC
rigidly](http://lists.llvm.org/pipermail/cfe-dev/2020-April/065271.html).

* ORC JIT Weekly #13 [is
out](http://lists.llvm.org/pipermail/llvm-dev/2020-April/141228.html).
Discord user @Juicebox has written a project to enable PDB debugging on
Windows with ORC and MCJIT.

* Chris Tetreault [provides an update on refactoring the class hierarchy of
VectorType](http://lists.llvm.org/pipermail/llvm-dev/2020-April/141162.html).

* Tom Stellard [shares the proposed 10.0.1 release
schedule](http://lists.llvm.org/pipermail/llvm-dev/2020-April/141128.html).
This would see rc1 tagged on May 18th, rc2 on June 18th, and 10.0.1-final on
June 25th.


## LLVM commits

* MC layer support for the Armv8.6-a matrix multiply extension was added.
[832cd74](https://reviews.llvm.org/rG832cd749131),
[e714683](https://reviews.llvm.org/rGe7146838800).

* `TargetTransformInfo::getOperationCost` was removed and merged into
getUserCost. [ee959dd](https://reviews.llvm.org/rGee959ddc5ee).

* The MachineVerifier is now run after more passes.
[46a52ff](https://reviews.llvm.org/rG46a52ff9eda).

* The integration of compiler extensions into the build system was updated.
[8f766e3](https://reviews.llvm.org/rG8f766e382b7).


## Clang commits

* The time profiler script will now emit a beginningOfTime parameter that can
be used for combining events from different processes while preserving their
relative timing. [2899103](https://reviews.llvm.org/rG2899103108d).

* More SVE builtins were added.
[823e2a6](https://reviews.llvm.org/rG823e2a670a9),
[db79974](https://reviews.llvm.org/rGdb7997472bb),
[c84e130](https://reviews.llvm.org/rGc84e1305c4f).


## Other project commits

* lld-macho gained basic support for linking against dylibs.
[060efd2](https://reviews.llvm.org/rG060efd24c7f).

* compiler-rt started to gain support for platforms with 16-bit int.
[4d41df6](https://reviews.llvm.org/rG4d41df64828).

* LLDB's reproducer infrastructure gained support for a new 'passive replay'
mode. [950a8aa](https://reviews.llvm.org/rG950a8aa165e).

* MLIR's liveness analysis now supports nested regions.
[c79227c](https://reviews.llvm.org/rGc79227cabb3).