Re: [creduce-bugs] clang pass bug

[John Regehr <regehr@cs.utah.edu>]

I hope that pastebin is more durable than the bug is :)

How hard is it to turn the bug into a regression test?  Do we have any 
infrastructure for clang-pass regression tests?  Shouldn't be too hard. 
In Souper we use lit (the LLVM test tool) and it is pretty decent.

John


On 1/24/18 8:06 AM, Eric Eide wrote:
> John Regehr <regehr@cs.utah.edu> writes:
>
> > Pascal Cuoq reports this one:
> > https://pastebin.com/D7aaZv5r
>
> Is pastebin durable?  (I don't know.)  If not, we should copy the test case
> someehere.