[llvm-dev] LLVM Weekly - #220, Mar 19th 2018
Alex Bradbury via llvm-dev
llvm-dev at lists.llvm.org
Mon Mar 19 15:29:40 PDT 2018
LLVM Weekly - #220, Mar 19th 2018
=================================
If you prefer, you can read a HTML version of this email at
<http://llvmweekly.org/issue/220>.
Welcome to the two hundred and twentieth 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
The LLVM Project Blog has a [new entry about
DragonFFI](http://blog.llvm.org/2018/03/dragonffi-ffijit-for-c-language-using.html),
which uses LLVM and Clang to provide a high quality foreign function interface
library for Python.
Tickets are [still
available](http://lists.llvm.org/pipermail/llvm-dev/2018-March/121783.html)
for the Women in Compilers and Tools reception, to be held the night before
the 2018 EuroLLVM Developers' Meeting (Sunday April 15th).
Bryce Lelbach has written a
[trip report](https://www.reddit.com/r/cpp/comments/854mu9/2018_jacksonville_iso_c_committee_reddit_trip/)
from the 2018 Jacksonville ISO C++ Commitee meeting.
## On the mailing lists
* Dimitry Andric has [collected
data](http://lists.llvm.org/pipermail/llvm-dev/2018-March/121871.html) on the
size of Clang executables and the time taken to compile Clang since
October 2015. Since then, total executable size grew by 43% and build time
increased by 60%.
* Guillaume Chatelet and Clement Courbet have posted an [RFC on
llvm-exegesis](http://lists.llvm.org/pipermail/llvm-dev/2018-March/121814.html),
a tool to automatically determine instruction scheduling properties (such as
latency and the number of micro-ops). Feedback has been very positive so far,
and Clement
[explained](http://lists.llvm.org/pipermail/llvm-dev/2018-March/121841.html)
hat work is underway to integrate exegesis with the recently contributed
llvm-mca.
* Tanya Lattner is [looking for more Birds of a Feather session
proposals](http://lists.llvm.org/pipermail/llvm-dev/2018-March/121782.html)
for EuroLLVM.
* Anastasia Stulova is seeking suggestions for [hacking sessions at EuroLLVM
this year](http://lists.llvm.org/pipermail/llvm-dev/2018-March/121850.html),
and also shares some initial ideas.
* Tom Stellard has proposed the [release chedule for 5.0.2 and
6.0.1](http://lists.llvm.org/pipermail/llvm-dev/2018-March/121757.html). The
5.0.2 release will contain only Spectre-related fixes. 6.0.1 will have its
first release candidate on April 15th, and aim for a final version on June
8th.
* Peter Smith has posted an RFC on [avoiding giving a default CPU to the LTO
plugin](http://lists.llvm.org/pipermail/llvm-dev/2018-March/121830.html). As
was
[determined](http://lists.llvm.org/pipermail/llvm-dev/2018-March/121845.html)
later in the thread, the issue is slightly wider in scope than just LTO.
* Dangeti Tharun Kumar is [seeking
feedback](http://lists.llvm.org/pipermail/llvm-dev/2018-March/121736.html) on
a proposal to add an LLVM front end for the P4 language (used for packet
forwarding).
* Artem Dergachev has offered an
[update](http://lists.llvm.org/pipermail/cfe-dev/2018-March/057255.html) on
efforts to improve handling of C++ temporary object construction and
destruction in the Clang static analyzer.
* Sylvestre Ledru
[proposes](http://lists.llvm.org/pipermail/llvm-dev/2018-March/121873.html)
reanming LLVM libraries and binaries to exclude the .0 suffix. e.g. clang-7
rather than clang-7.0.
* James Henderson is
[proposing](http://lists.llvm.org/pipermail/llvm-dev/2018-March/121727.html)
to update the version of googletest imported into the LLVM tree to a
non-release tagged version. This is because the last tagged release occurred
in August 2016, there have been desirable changes since then, and there is no
clear timeline on a new upstream release.
* Scott Linder has posted an RFC on [adding a MessagePack reader/write to
LLVM](http://lists.llvm.org/pipermail/llvm-dev/2018-March/121740.html).
## LLVM commits
* lit has been extended to support reporting multiple 'micro-test' results, to
enable it to be used for collecting microbenchmark results.
[r327422](http://reviews.llvm.org/rL327422).
* The handling of dangling debug info in the SelectionDAG has been improved.
[r327303](http://reviews.llvm.org/rL327303).
* The LoopInstSimplify pass has been removed due to the fact that it is not
used in-tree and has no tests. [r327329](http://reviews.llvm.org/rL327329).
* dsymutil will now perform `analyzeContextInfo` and `CloneDIEs` in parallel,
leading to performance improvements of up to ~30%.
[r327399](http://reviews.llvm.org/rL327399).
* `!dag` has been added to Tablegen, enabling programmatic construction of DAG
nodes. [r327492](http://reviews.llvm.org/rL327492).
* FastISel will now sink the materialisation of local values to their first
use. [r327581](http://reviews.llvm.org/rL327581).
* llvm-dsymutil has been renamed to dsymutil.
[r327790](http://reviews.llvm.org/rL327790).
## Clang commits
* The Clang C++ status page has been updated.
[r327769](http://reviews.llvm.org/rL327769).
* clang-tidy gained a Zircon module (Zircon forms the core of Fuchsia OS). It
includes checks specific to the Zircon kernel.
[r327590](http://reviews.llvm.org/rL327590).
## Other project commits
* Prologue handling in LLDB has been improved, adding support for functions
with multiple entry points. [r327331](http://reviews.llvm.org/rL327331).
* lld gained support for the `--cref` option to print out a table of symbols
and filenames. [r327565](http://reviews.llvm.org/rL327565).
More information about the llvm-dev
mailing list