A little later this week than usual, but here's what's happened over the past two weeks!
Friend of the Tree
@Gankra has nominated @tshepang for Friend of the Tree this week:
Over the last year Tshepang has landed over 100 improvements to our documentation. Tshepang saw where documentation was not, and said "No. This will not do."
We should all endeavor to care about docs as much as Tshepang.
Compiler
We decided to accept RFC #1191, on adding a HIR to the compiler.
In addition, RFC #1211 has entered into final comment period. This RFC describes a mid-level, simplified IR that will be constructed after type-checking but be used to drive advanced analysis and code generation.
-- nmatsakis
Language
Two decisions were made since the last report:
- RFC #873 (macros in types) was accepted.
- RFC #1150 (re-exporting associated iteams with
use
) was withdrawn.
One RFC is entering final comment period:
- RFC #1219: extending the
use
syntax to allow renaming in group imports. - RFC #1214: clarifying rules for projections and well-formedness. This is a soundness RFC that clarifies some of the language rules, closing various soundness issues in the process. Some existing code may no longer compile (full evaluation in progress); in any case, the plan is to phase the changes in gradually.
Finally, we would like to call attention to one pending RFCs:
- RFC #1177: introducing an anonymous placeholder lifetime
'_
. There are some debates as to the best defaulting behavior and best syntax; more input would be helpful.
Thanks!
--- nmatsakis
Libs
A little belated this week and we slipped last week, so there's definitely a few things to catch up on! The libs team didn't hold a triage meeting this week due to many attendees having conflicts at the normal time, but we should make progress on the FCP RFCs and new PRs next week. As a summary of last week's actions:
RFCs going into their final comment period:
- FCP PR #1184: RFC: Stabilize the #![no_std] attribute
- FCP PR #1183: RFC: Allow changing the default allocator
- FCP PR #770: io error handling design
Decisions from last week:
- FCP PR #1174:
RFC for creation of
IntoRaw{Fd, Socket, Handle}
trait to complementAsRaw*
- PR #26743:
Add a method
lines_any
toBufRead
- Closed. The associated RFC will be considered instead.
- FCP PR #1158:
RFC: Expand the std::net module
- Postponed, pending some policy decisions about how to handle major new API additions in std. We should have more details on this next week.
In the triage meeting last week we also discussed the story for adding new large
portions of functionality to the standard library (e.g. the std::net
RFC), and
we will be posting our conclusions soon about this topic.
Tools
It's been a bit since the last report, so here's some of the highlights over the past two weeks:
- 32-bit MSVC nightlies are now available
- The
--cap-lints
RFC was merged and a PR was posted - Travis CI is now used more aggressively on the main repo thanks to @Gankra
- @vadimcn has done some great work to help bring unwinding to 64-bit MSVC soon.
- @michaelwoerister has gotten MSVC debuginfo working
- LLVM was updated to enable 32-bit MSVC to bootstrap and LLVM's
archive support is leveraged on more platforms instead of using the system
ar
. - @dhuseby has taken over the FreeBSD bots and snapshots will soon be registered to get them green again!
In other news the cargo install
RFC is likely to go into FCP
soon, so be sure to keep an eye on it!