On Apr 29, 2009, at 1:53 PM, Eli Friedman wrote: > On Wed, Apr 29, 2009 at 10:35 AM, steve naroff <[email protected]> > wrote: >> Another point...the ObjC rewriter project is still considered >> experimental >> (since the generated code isn't compiler or target neutral). > > I'm not sure the reasoning makes sense, but I agree it doesn't need to > be held to the same standards as other code.
My point was this...investing the time/effort in developing *meaningful* rewriter tests for one target/compiler hasn't been a priority (given other clang projects and the experimental nature of this work). btw...the existing tests in test/Rewriter are mostly front- end sanity checks...they don't test for runtime correctness. > Testcases are always nice, though. > Absolutely. My commits to clang "proper" almost always contain a test case. When I forget, it's a bug (and I'm happy to add one). In this instance, it would be too much work for me to do this retroactively (with little payoff, given my previous comments). snaroff >> Nevertheless, we do have clients who are testing these changes on a >> regular >> basis (which gives us some regression coverage). > > Okay, that's good to know. > > -Eli _______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
