Backport an LLVM fix to close #9117#9118
Conversation
|
Doesn't this need llvm-auto-clean-trigger? |
|
Ah yes, it does indeed (updated now) |
|
@alexcrichton: could we try updating to the latest LLVM if we're going to have a rebuild anyway? |
|
I'm hesitant to update LLVM when we're still dealing with minor fallback from the last upgrade. It's only been a month and I don't think that anything earth-shattering has happened in LLVM (although I may be wrong?). We're not adding any extra cruft into our fork (we'll just have to remember to strip the patch before rebasing), so for now I'd be in favor of just bringing this patch in instead of upgrading llvm entirely again. |
|
This is getting some weird permission error build problems on windows, and in the past I've seen that if you just re-build llvm these problems go away. I'm waiting for an empty queue and then I'm going to retry this twice in a row (hopefully) |
|
@alexcrichton: I just think it would be easier to roll a bit more with LLVM (upgrade once a month) and deal with issues as they come up. It's much more frustrating to tackle a bunch at once. In general, any upgrade is going to fix more bugs than it introduces. |
|
Worth the trouble at this point, just rebased on llvm-mirror/master and am rebuilding (will push/re-title once finished). |
I created a new branch rust-llvm-2013-09-11 in my llvm repo to mark this momentous occasion
I created a new branch rust-llvm-2013-09-11 in my llvm repo to mark this momentous occasion