LLVM Weekly - #135, Aug 1st 2016
Welcome to the one hundred and thirty-fifth 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. Subscribe to future issues at https://llvmweekly.org and pass it on to anyone else you think may be interested. Please send any tips or feedback to asb@asbradbury.org, or @llvmweekly or @asbradbury on Twitter.
News and articles from around the web
LLVM/Clang 3.9.0-rc1 has been tagged. It's time to get testing.
Swift 3 is nearing completion and Chris Lattner has shared some thoughts looking ahead to Swift 4.
The first LLVM Cauldron will be held on September 8th in Hebden Bridge, UK. Registration to attend is open, and you have a week to submit your talks. Also consider the 2016 US LLVM Developers' Meeting, to be held November 3rd-4th in San Jose, California.
On the mailing lists
-
Renato Golin has shared a proposed LLVM target acceptance policy.
-
The discussion about a potential move from SVN to Git continues with most of it focused on the pros and cons of a single monolothic repo for all LLVM subprojects. Mehdi Amini has really helpfully collected links to various workflows people have shared and other concrete reasoning for or against monorepos that people have shared. Chris Bieneman has shared his concerns about a monorepo and is keen to point out that lack of dissent in this particular thread doesn't mean consensus - as others have pointed out, the discussion has become rather too unwieldy for many people to follow.
-
Christopher Bergström is interested in organising an LLVM social in Asia and want to gauge interest.
-
Michael Lewis wonders if there's interest in generating PDB from LLVM-backed languages. The reason to ask is actually as he's done rather a lot of work on it for his own language and wonders if others may benefit. A number of responders suggest looking at the PDB writing support that is being done in-tree.
LLVM commits
-
A massive, target-independent MachinePipeliner pass has landed which implements Swing Modulo Scheduling. r277169.
-
bugpoint has been modified to use a simpler control-flow graph simplification routine that results in faster runtime and more minimal test cases. r277063.
-
llvm-cov learned to export coverage data to JSON. r276813.
-
GlobalISel now has an instruction selector. r276875.
Clang commits
-
A simple analyzer was added to detect copy and pasted code. r276782.
-
The web manual for Clang checker developers has been updated with a much expanded section on debugging. r277029.