Remove length check from ThinSlice::clone to halve generated code

[arthurprs]

Remove length check from ThinSlice::clone. This check doubles the size of the function generated code. Since ThinSlice::clone is inlined everywhere this can improve codegen significantly.

Includes a commit bumping the version to 0.1.10.

Asm before

        .cfi_startproc
        sub rsp, 56
        .cfi_def_cfa_offset 64

        mov rax, qword ptr [rdi]

        mov rcx, qword ptr [rax + 8]

        lock inc        qword ptr [rax]

        jle .LBB366_3

        mov qword ptr [rsp], rcx

        cmp qword ptr [rax + 8], rcx

        jne .LBB366_4

        add rsp, 56
        .cfi_def_cfa_offset 8
        ret

.LBB366_3:
        .cfi_def_cfa_offset 64
        call qword ptr [rip + std::process::abort@GOTPCREL]

        ud2

.LBB366_4:
        lea rcx, [rip + .L__unnamed_289]

        mov qword ptr [rsp + 8], rcx
        mov qword ptr [rsp + 16], 1
        mov qword ptr [rsp + 24], 0
        lea rcx, [rip + .L__unnamed_183]
        mov qword ptr [rsp + 40], rcx
        mov qword ptr [rsp + 48], 0
        add rax, 8

        mov rsi, rsp
        lea rdx, [rsp + 8]
        mov rdi, rax

        call core::panicking::assert_failed
        ud2

[Manishearth]

Thanks!

[Manishearth]

Ugh, oops, I made a mistake here, the 0.1.10 bump was already on my local master. I'm going to force push to fix this and repush your commit

[Manishearth]

Pushed 4b219e3

[Manishearth]

Pushed 4b219e3. Make a separate PR for 0.1.11 if you need it.