about summary refs log tree commit diff
path: root/src/rustllvm/PassWrapper.cpp
diff options
context:
space:
mode:
authorbors <bors@rust-lang.org>2014-04-28 05:21:46 -0700
committerbors <bors@rust-lang.org>2014-04-28 05:21:46 -0700
commita1ad41b93d133aa4f3bda71475f8e41d9dfe704d (patch)
treea8cdfdaf2359e39d80de4045ca7f1143f7767b43 /src/rustllvm/PassWrapper.cpp
parent7a19a82d119ca51ed872ed207bed396cdf4a3283 (diff)
parentc8a29c4c595e76b71372a2e40d359ac1ddd8aec8 (diff)
downloadrust-a1ad41b93d133aa4f3bda71475f8e41d9dfe704d.tar.gz
rust-a1ad41b93d133aa4f3bda71475f8e41d9dfe704d.zip
auto merge of #13791 : lifthrasiir/rust/mod-inner-span, r=huonw
This PR is primarily motivated by (and fixes) #12926.

We currently only have a span for the individual item itself and not for the referred contents. This normally does not cause a problem since both are located in the same file; it *is* possible that the contained statement or item is located in the other file (the syntax extension can do that), but even in that case the syntax extension should be located in the same file as the item. The module item (i.e. `mod foo;`) is the only exception here, and thus warrants a special treatment.

Rustdoc would now distinguish `mod foo;` from `mod foo {...}` by checking if the span for the module item and module contents is in different files. If it's the case, we'd prefer module contents over module item. There are alternative strategies, but as noted above we will have some corner cases if we don't record the contents span explicitly.
Diffstat (limited to 'src/rustllvm/PassWrapper.cpp')
0 files changed, 0 insertions, 0 deletions