Subteam reports 2016-01-15

[aturon]

Lang

Full report

Three RFCs went into final comment period this week:

• RFC 243, which adds a ? operator that is roughly equivalent to the existing try! macro, except that it operates in suffix position. The scope of the ? operator can be controlled using a try { ... } block and there is also sugar for handling errors try { .. } catch { .. }. This RFC has been around since before 1.0, and has seen quite a lot of discussion. (Note that the try and catch keywords are likely to change, given the overlap with the existing try! macro.
• RFC 1320 which modifies the RangeInclusive type from a struct to an enum.
• RFC 1462, which adds the open brace character to the follow set for type fragments. This macro is taking a “fast path” to FCP due to the time-sensitive nature of its contents – it is intended to mitigate the effects of a recent bugfix.

We are seeking feedback on the name and syntax for so-called OIBIT traits, such as Send and Sync.

Finally, we’ve decided to accept RFC 1331, which spells out a procedure for making an official Rust grammar that is independent from the implementation.

Libs

Full report

This week, we made decisions on all the outstanding FCP issues for Rust 1.7:

• FCP Issue #23284: Tracking issue for Path::relative_from stabilization
  • Stabilize (with renaming and slight type change)
• FCP Issue #27709: Tracking issue for Ipv{4,6}Addr convenience methods
  • Stabilize subset of methods confirmed to match official spec
• FCP Issue #27712: Tracking issue for fs_mode stabilization
  • Deprecate in favor of libc
• FCP Issue #27713: Tracking issue for custom hashers in HashMap
  • Stabilize with renaming
• FCP Issue #27722: Tracking issue for path_prefix stabilization
  • Deprecate
• FCP Issue #27728: Tracking issue for formatting with a radix
  • Deprecate
• FCP Issue #27729: Tracking issue for as_slice stabilization
  • Stabilize
• FCP Issue #27735: Tracking issue for IntoCow stabilization
  • Deprecate
• FCP Issue #27739: Tracking issue for Zero/One/iter_arith stabilization
  • Punt
• FCP Issue #27750: Tracking issue for clone_from_slice stabilization
  • Stabilize with new panic behavior
• FCP Issue #27753: Tracking issue for num_bits_bytes stabilization
  • Deprecate
• FCP Issue #27755: Tracking issue for wrapping stabilization
  • Stabilize inherent methods
• FCP Issue #27782: Tracking issue for DebugTuple::formatter
  • Deprecate
• FCP Issue #27791: Tracking issue for str escaping
  • Punt
• FCP Issue #27798: Tracking issue for sync::Semaphore
  • Deprecate
• FCP Issue #28796: Tracking issue for FnBox()
  • Punt
• FCP Issue #29157: Tracking issue for owned CString conversions
  • Stabilize
• FCP Issue #30423: Tracking issue for slice::sort_by_key
  • Stabilize

Next week we return to our regular programming!

Compiler

Full report

We have decided to place RFC 1457 into final comment period. This RFC concerns the way that for loops are internally implemented in the compiler.

Tools

Full report

• Initial support for PNaCl is in the Makefiles
• The compiler itself now has rpath enabled by default to avoid requiring the usage of LD_LIBRARY_PATH for downloaded installers.
• A patch to use the gold linker by default landed, but it was unfortunately reverted due to bugs
• The compiler can now emit errors in JSON format
• Initial support for powerpc64 has landed.
• The website has seen a major revamp.
• The RFC for Cargo target specific dependencies will remain in FCP for awhile longer as the syntax is hammered out a bit more.