From: Christian Collberg <collberg@gm...> - 2012-06-18 01:16:23
I am in need of CIL to be extended to handle gcc's
labels-as-values better. Currently they're compiled
out to switch statements, but I need an option to
turn off this behavior. Instead, there needs to be
the appropriate new abstract syntax added to
CIL to manipulate indirect gotos, local &&labels,
arrays of such labels, etc.
I have $s allocated for this project.
If you're interested, please send me an outline
of the tasks that you believe need to be performed,
your timeline (sooner is better than later!), your
background in working with CIL, and how much
you will charge.
I'm also interested to hear from the community
if you would like to see this option folded into
CIL's mainline code.
Department of Computer Science
University of Arizona
sɐpıɯ ƃuıʞ ǝsɹǝʌuı
From: Gabriel Kerneis <gabriel@ke...> - 2012-06-18 06:59:35
On Mon, Jun 18, 2012 at 09:15:57AM +0800, Christian Collberg wrote:
> I'm also interested to hear from the community
> if you would like to see this option folded into
> CIL's mainline code.
This is an issue similar to the Question variant introduced in CIL 1.5.0. It
might break existing programs (non-exhaustive pattern-matching), but as long as
it is disabled by default, I see no reason not to include it. Feel free to
discuss design decisions on this list.
Get latest updates about Open Source Projects, Conferences and News.