[LLVMdev] type-system-rewrite branch near landing
clattner at apple.com
Thu Jul 7 23:30:58 PDT 2011
On Jul 7, 2011, at 2:09 PM, Jay Foad wrote:
>>> 1. Many Objective-C(++) tests fail, because they use implicitly
>>> defined structs for various ObjC runtime data structures; the
>>> ASTConsumer HandleTagDeclDefinition callback is never called for these
>>> structs, which means that I don't bother to lay them out, so they only
>>> ever get an opaque LLVM type. Then we get assertion failures when
>>> generating code to access fields in these structs.
>> Is there some reason we can't just layout types lazily when requested,
>> like the old code does? Should save some work in common cases as well
>> as implicitly solve any issues here.
> I couldn't get it working that way, but I can't remember exactly what
> the problems were. I'll have another go. It would certainly be less
> disruptive if it still worked like that.
FYI, I merged mainline up to r134683 into the branch, and fixed the last llvm-test failures. I'm good to land it now, except for the small matter of clang/llvm-gcc/dragonegg. :)
I'm going to spend tomorrow tackling llvm-gcc. Once I'm done with it I'll take a look at Clang if you haven't beaten me to it.
If no one is interested in updating dragonegg, then I'll have to land the branch without updating it. It should be relatively easy to update dragonegg once llvm-gcc is updated. Worst case, its build can be fixed once Duncan gets back from vacation.
More information about the llvm-dev