Subteam reports 2016-01-08


#1

Libs

Full report

This week was devoted entirely to triage catch up: we discussed a large number of pending issues and PRs and issues decisions on them.

Next week we will reach final decisions on those features in final comment period for stabilization in 1.7.

Lang

Full report

We decided to move RFC #1331 into FCP. This RFC proposes to start the development of an official Rust grammar and lays out a plan to get there. (The RFC does not propose to autogenerate the Rust parser from the grammar; that’s an implementation detail in any case, better suited to a compiler team RFC.) This is not expected to have any direct impact on end-users, but is an important step towards formalizing Rust’s definition and specification.

We also recently discovered a flaw in the macro future-proofing RFC (issue #30531). This flaw means that otherwise legal changes to the expression grammar (for example, the addition of type ascription) can break macros “in the wild”. We intend to close the flaw, but in the meantime macro authors may want to review their macros to see if they will be affected by upcoming grammar changes, such as type ascription or the possible addition of a ? operator. See this comment for more details.

Compiler

Full report

This week has seen a number of MIR PRs land, including a large overhaul of how calls and panics are handled by Simonas Kazlauskas, a critical bug fix by Michael Woerister, and new and improved MIR dumps in both graphviz and textual format by Scott Olson. Nice work!

There is continued progress towards incremental compilation. In particular, PR #30532 (by yours truly) lays out the infrastructure for tracking dependencies within the compiler.