Uh oh!
There was an error while loading. Please reload this page.
Move gensym operations from Symbol to Ident - #60903
Conversation
There was a problem hiding this comment.
| /// constructs a new identifier from a symbol and a span. | |
| /// Constructs a new identifier from a symbol and a span. |
petrochenkov
commented
May 17, 2019
r=me after rebase |
17b5966 to
ad5dd49Comparennethercote
commented
May 19, 2019
@bors try |
bors
commented
May 19, 2019
…=<try> Move gensym operations from `Symbol` to `Ident` Gensyms are always at the `Ident` level, and long-term we probably want to record gensym-ness in hygiene data. r? @petrochenkov
ad5dd49 to
8a96f26Comparebors
commented
May 20, 2019
☔ The latest upstream changes (presumably #60969) made this pull request unmergeable. Please resolve the merge conflicts. |
Note that the `is_gensymed` call on `primitive_types` is unnecessary because that table only contains the name of primitive types (e.g. `i32`) and never contains gensyms.
8a96f26 to
88d2999Comparennethercote
commented
May 20, 2019
@bors try |
bors
commented
May 20, 2019
…=<try> Move gensym operations from `Symbol` to `Ident` Gensyms are always at the `Ident` level, and long-term we probably want to record gensym-ness in hygiene data. r? @petrochenkov
bors
commented
May 20, 2019
☀️ Try build successful - checks-travis |
nnethercote
commented
May 20, 2019
@rust-timer build e67ebe6 |
rust-timer
commented
May 20, 2019
Success: Queued e67ebe6 with parent 128b4c8, comparison URL. |
rust-timer
commented
May 20, 2019
Finished benchmarking try commit e67ebe6: comparison url |
nnethercote
commented
May 21, 2019
It's a slight performance regression, but I think it's necessary to take at some point in order to move gensym-ness away from symbols. We've had a bunch of significantly larger performance improvements related to symbol handling recently, which more than make up for this. @bors r=petrochenkov |
bors
commented
May 21, 2019
📌 Commit 88d2999 has been approved by |
…=petrochenkov Move gensym operations from `Symbol` to `Ident` Gensyms are always at the `Ident` level, and long-term we probably want to record gensym-ness in hygiene data. r? @petrochenkov
bors
commented
May 21, 2019
bors
commented
May 21, 2019
☀️ Test successful - checks-travis, status-appveyor |
Gensyms are always at the
Identlevel, and long-term we probably want to record gensym-ness in hygiene data.r? @petrochenkov