>>! In D4759#6, @EricWF wrote:
> As much as I would like to take testit out back and shoot it, previous 
> conversations with @mclow.lists leads me to believe its still widely used. 
> Since lit usually isn't available outside of the LLVM repo I'm not sure that 
> removing testit is viable at this point. 
> 
> I would like to see LIT become a standalone "first-class" binary so it could 
> be easier to use outside of makefile rules. Daniel Dunbar expressed some 
> thoughts to this end a while ago. I will follow those up with him.
> 
> My current workflow is something like:
> ```
> alias lit="python /path/to/llvm/utils/lit/lit.py"
> cp libcxx-build/lit.site.cfg libcxx/test/lit.site.cfg
> ```
> 
> Until its as easy to use LIT in the source tree as it is the testit script I 
> don't think we can even consider removing it.
>  

I wonder how much of the reason that testit is more widely used is because it's 
the most obvious path to running the tests. I know when I first started working 
with the project I had no idea that LIT was even there (or what it was, for 
that matter). The libc++ docs make no mention of how to use LIT for the tests 
(by contrast, it's now the only thing mentioned for libc++abi, albeit 
indirectly though the cmake target).

Still, I'll concede that removing testit isn't something we can do just yet.

http://reviews.llvm.org/D4759



_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits

Reply via email to