Uh oh!
There was an error while loading. Please reload this page.
debuginfo: Fix type description generic enum discriminants. - #27070
Conversation
rust-highfive
commented
Jul 16, 2015
r? @Aatch (rust_highfive has picked a reviewer for you, use r? to override) |
alexcrichton
commented
Jul 16, 2015
tamird
commented
Jul 16, 2015
Something wrong with this sentence |
Manishearth
commented
Jul 17, 2015
This might be causing a test failure http://buildbot.rust-lang.org/builders/auto-mac-64-opt/builds/5712 |
bors
commented
Jul 17, 2015
⌛ Testing commit f9a20bb with merge b465213... |
Manishearth
commented
Jul 17, 2015
@bors r- |
Manishearth
commented
Jul 17, 2015
@bors r- force (fails tests) |
alexcrichton
commented
Jul 17, 2015
I'm gonna land this as part of #27076 and just ignore the LLDB half of things for now (this is being tested on gdb and fixes bugs elsewhere in the upgrade). I'll open an issue about dealing with the LLDB failure, however, so we can be sure to keep track of it as it definitely looks like it should be investigated! |
Manishearth
commented
Jul 17, 2015
Sure! (was just r-ing so it won't clog up the queue 😄 ) |
This fixes an issue that @alexcrichton found while trying to upgrade LLVM: So far the code has assumed that the type of a generic enum's discriminant always was the same, regardless of any type arguments. This is not true. An
Option<i16>will have a 16 bit discriminant, while anOption<i32>will have a 32 bit discriminant. This PR fixes this.