On Sat, Sep 8, 2012 at 10:03 AM, Anton Korobeynikov <[email protected]> wrote: >> How'd you come across this? Any reasonable test case? > Well... Just release w/o asserts mode on darwin + make test. > >> (alternatively/in addition: should we annotate StringRef's ctor with >> nonnull? (I don't suppose we bother with this in general, but given >> the prevailing LLVM style of using raw zero literals (rather than the >> NULL macro) for null pointers, I wonder whether it'd help catch a few >> surprising cases)) > There is an assert inside stringref ctor, this is how it was caught :)
Oh, OK - sounds good :) _______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
