Library FCP issues for 1.7

[alexcrichton]

The set of issues for stabilization in the 1.7 release have now entered their final comment period. This means that over the next cycle we’ll discuss the issues and decide on whether they should be stabilized or deprecated for the 1.7 release (2016/3/3). As a reminder, if anything major changes about an API it will be punted out of FCP for another cycle (whether it’s being deprecated or stabilized).

This list is a bit shorter than the previous releases as we’re winnowing down the unstable features in the standard library and the holidays are also just around the corner, but here’s the list!

• path::relative_from - stabilize after renaming
• ip methods - stabilize after investigation
• fs_mode - deprecate
• custom hashers - stabilize with a renaming
• path_prefix - deprecate
• formatting w/ radix - deprecate
• as_slice - deprecate
• IntoCow - deprecate
• iter_arith - stabilize
• clone_from_slice - stabilize after deciding panic semantics
• num_bits_bytes - deprecate
• wrapping - stabilize in a different location
• debug tuple formatter - deprecate
• str escaping - stabilize
• Semaphore - deprecate
• fnbox - stabilize if coherence is appeased
• cstring owned conversions - stabilize
• sort_by_key - stabilize

Note that issues often have more detailed or nuanced comments about stabilization or deprecation on the issue itself, so if something seems surprising be sure to click through and read the discussion. Also feel free to contribute to any discussion!

[ubsan]

I’m not sure why clone_from_slice is called clone_from_slice. It has the exact same semantics as clone_from, it should have the same name (especially since the current name is really long and unwieldy). The only reason clone_from_slice is a thing is because we can’t implement Clone on [T], so it seems silly to name it something different than the method it… is.