LLVM Weekly - #18, May 5th 2014
==============================
If you prefer, you can read a HTML version of this email at
<http://llvmweekly.org/issue/18>
Welcome to the eighteenth 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](http://asbradbury.org).
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.
I'm going to be in the San Francisco area May 13th-20th with some other
Raspberry Pi people. We'll be at Maker Faire Bay Area on the 17th and 18th.
Let me know if there's anything else I should check out while over there.
## News and articles from around the web
Andrew Ruef has written a blog post about [using static analysis and Clang to
find the SSL heartbleed
bug](http://blog.trailofbits.com/2014/04/27/using-static-analysis-and-clang-to-find-heartbleed/).
The code for the checker described in the blog post is [available on
Github](https://github.com/awruef/find-heartbleed).
The FTL ('Fourth tier LLVM') Javascript JIT is [now enabled in WebKit
for
Mac](https://trac.webkit.org/changeset/167958). The WebKit Wiki has [more
information](https://trac.webkit.org/wiki/FTLJIT). I haven't seen any public
benchmark figures. Please do share if you have any.
Eli Bendersky has written an article about [how to use libTooling to implement
source to source
transformations](http://eli.thegreenplace.net/2014/05/01/modern-source-to-source-transformation-with-clang-and-libtooling/).
The next [Paris LLVM
Social](http://article.gmane.org/gmane.comp.debugging.lldb.devel/3828) will
take place on May 5th (i.e. this evening).
The LLVM Bay Area social [will take place on May
8th](http://article.gmane.org/gmane.comp.compilers.clang.devel/36443). Please
RSVP if you are interested.
## On the mailing lists
* Discussion is still going on regarding [last week's proposal to add
intrinsics for safe
division](http://article.gmane.org/gmane.comp.compilers.llvm.devel/72466).
Philip Reames has [shared his reasoning againt the
proposal](http://article.gmane.org/gmane.comp.compilers.llvm.devel/72588).
This triggered a *lot* of discussion which is unfortunately more than I have
time to summarise.
* Kevin Modzelewski is working on exceptions for Pyston, the LLVM-based Python
JIT and [asks for advice/experiences on how others have implemented them in
projects like
VMKit](http://article.gmane.org/gmane.comp.compilers.llvm.devel/72668). There
are a number of useful responses.
* Alp Toker [suggests adding a coding style policy to discourage excessive use
of default
arguments](http://article.gmane.org/gmane.comp.compilers.clang.scm/97880).
* Yi Kong has [suggested improvements to LNT for
benchmarking](http://article.gmane.org/gmane.comp.compilers.llvm.devel/72598).
One of the concerns raised in followup responses is that benchmark results may
not be stable and may take too long to be practical on e.g. ARM hardware.
Chris Matthews suggests [performing a geometric mean similar to the SPEC
benchmarks](http://article.gmane.org/gmane.comp.compilers.llvm.devel/72619) as
an alternative to a plain total execution time metric.
* Dario Domizioli has [posted an RFC for an optnone
pragma](http://article.gmane.org/gmane.comp.compilers.clang.devel/36314). This
proposal triggered robust discussion about whether there is any advantage to
supporting this in addition to the optnone function attribute, though it seems
as though there is enough support for the feature for it to make its way in.
## LLVM commits
* The patch to [perform common subexpression elimination for a group of
getelementptrs](http://article.gmane.org/gmane.comp.compilers.llvm.devel/72366)
that was discussed a couple of weeks ago has been merged. It is currently only
enabled for the NVPTX backend. [r207783](http://reviews.llvm.org/rL207783).
* X86 code generation has been implemented for the musttail function
attribute. [r207598](http://reviews.llvm.org/rL207598).
* Pass run listeners were added to the pass manager. This adds C/C++ APIs to
enable fine-grain progress report and safe suspension points. See the commit
message for more info [r207430](http://reviews.llvm.org/rL207430).
* The optimisation remark system has started to be used, with calls to
emitOptimizationRemark added to the loop unroller and vectorizer.
[r207528](http://reviews.llvm.org/rL207528),
[r207574](http://reviews.llvm.org/rL207574).
* The SLPVectorizer gained the ability to recognize and vectorize intrinsic
math functions. [r207901](http://reviews.llvm.org/rL207901).
## Clang commits
* NRVO (named return value optimisation) determination was rewritten.
According to the commit message, "a variable now has NRVO applied if and
only
if every return statement in that scope returns that variable." Also, NRVO
is
performed roughly 7% more often in a bootstrap clang build.
[r207890](http://reviews.llvm.org/rL207890).
* libclang's documentation comment API has been split in to a separate
header.
[r207932](http://reviews.llvm.org/rL207932).
* The SLPVectorizer (superword-level parallelism) is now disabled at O0, O1
and Oz. [r207433](http://reviews.llvm.org/rL207433). It was later re-enabled
at Oz. [r207858](http://reviews.llvm.org/rL207858).
* The libclang API now supports attributes 'pure', 'const', and
'noduplicate'.
[r207767](http://reviews.llvm.org/rL207767).
* The comment parser no longer attempts to validate HTML attributes (the
previous solution was insufficient).
[r207712](http://reviews.llvm.org/rL207712).
## Other project commits
* `R_MIPS_REL32` relocation are now supported in lld.
[r207494](http://reviews.llvm.org/rL207494).
* A collection of CTRL+C related issues were fixed in lldb.
[r207816](http://reviews.llvm.org/rL207816).