[llvm-dev] Usage of LLJIT & Orc v2 by new end-user

[Daniel Chapiesky via llvm-dev]

Hello,

LLJIT and Orc v2 represent, to me, a great advancement for use of JIT in my
projects where I need granular control of practically everything.

Having lost my mind with glee after watching "Updating ORC JIT for
Concurrency", I eagerly downloaded the latest git version to examine the
LLJIT code.

Mentioned at the end of the video was that documentation was being worked
on and would be forthcoming.

Question 1: Is the draft documentation for LLJIT/Orc v2 available for
review even in its unfinished state... I ask because I need it... like bad.

Learning from code examples as in lli.cpp as well as the doxygen docs is
all well and good, but ancillary docs from the author would be invaluable,
even in unfinished form.

I volunteer to review and use this documentation as an unwashed noob to Orc
v2/LLJIT and give my best take on clarity and robustness.

The video showed a tool which was awesome.

The most important thing it demonstrated, again to me, was granular events
being emitted at each point in the JIT process...  these event messages are
what made me jump for joy...

However, looking at LLJIT, I am at a loss as to how to implement this kind
of event messaging myself....  to be honest, I got lost understanding the
error failure system and am still working on becoming proficient in its
use....

Question 2: Are the event emitters there and I am being daft? Or were the
events depicted in the video part of a code base that is yet to be released
(if ever)?

I will continue to enthusiastically explore and work with LLJIT, but any
help in regards to additional documentation would be greatly appreciated.

Excellent work so far.

Cheers,
Daniel Chapiesky
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20190330/9a1a263b/attachment.html>