Uh oh!
There was an error while loading. Please reload this page.
resolve: Get rid of try_define - #31308
Conversation
jseyfried
commented
Jan 31, 2016
@petrochenkov Nice! Looks good to me. |
jseyfried
commented
Feb 1, 2016
nikomatsakis
commented
Feb 2, 2016
/me waits |
bors
commented
Feb 11, 2016
☔ The latest upstream changes (presumably #31461) made this pull request unmergeable. Please resolve the merge conflicts. |
nikomatsakis
commented
Feb 13, 2016
Given the decision on #31337 -- should we close this PR? |
petrochenkov
commented
Feb 13, 2016
@nikomatsakis |
jseyfried
commented
Feb 13, 2016
Items that are partially shadowed (for example, a tuple struct shadowed by a function) currently have to be exported to expose the unshadowed namespace. This PR would still cause duplicate errors in that case. |
petrochenkov
commented
Feb 13, 2016
Hm, indeed. Closing this then. |
This was an unresolved question during review of #30843.
Module reexport sets were not deduplicated, so variant reexports ended up in pairs in these sets, and then in metadata, and then led to "duplicate definition" errors unless duplicate checking were turned off.
cc @jseyfried
r? @nikomatsakis