Rename get_bigfloat_precision, set_bigfloat_precision and with_bigfloat_precision

[dpsanders]

The function names get_bigfloat_precision, set_bigfloat_precision and with_bigfloat_precision are cumbersome and un-Julian.
I suggest changing these as follows:

• get_bigfloat_precision() -> precision(BigFloat)
• set_bigfloat_precision(n) -> set_precision!(BigFloat, n) or just set_precision!(n) or precision!(n)
• with_bigfloat_precision(n) -> with_precision(n)

Cc. @simonbyrne @stevengj @jiahao

[ScottPJones]

Please! Of course, this also works for decimal float, maybe unums.

[dpsanders]

@ScottPJones Yes, good point.

[ScottPJones]

You might need with_precision(T, n) as well as with_precision(n), to be able to handle setting precision for BigFloat, decimal floats, unums [when we get them]) separately, same as set_precision!

[tkelman]

probably drop the underscores, readable enough as setprecision! and withprecision IMO

[dpsanders]

@tkelman I partially agree, but currently set_rounding is written with an underscore (but without a !...)

set_bigfloat_precision, set_rounding and set_zero_subnormals are the only so-named functions in Base. Perhaps it's time to remove the set_ and just use !?

[ScottPJones]

The underscores do make it easier to read, there have even been studies on that (comparing camel case to separation with underscores to no separation).
I think having a set_XXX(...) is easier to read if you have a XXX(...), rather than XXX!(...).
Also, I'm not even sure it should have the !, because it is not modifying any of the arguments, it is modifying some state.

[StefanKarpinski]

Studies or no, we don't use underscores on Base functions.

[ScottPJones]

set_rounding???

[ScottPJones]

I honestly have a hard time believing you said that.

A_ldiv_B!, A_ldiv_Bc, A_ldiv_Bt, A_mul_B!, A_mul_Bc, A_mul_Bc!, A_mul_Bt, A_mul_Bt!, A_rdiv_Bc, A_rdiv_Bt,
Ac_ldiv_B, Ac_ldiv_Bc, Ac_mul_B, Ac_mul_B!, Ac_mul_Bc, Ac_mul_Bc!, Ac_rdiv_B, Ac_rdiv_Bc, At_ldiv_B, At_ldiv_Bt,
At_mul_B, At_mul_B!, At_mul_Bt, At_mul_Bt!, At_rdiv_B, At_rdiv_Bt,
count_ones, count_zeros, leading_ones, leading_zeros, trailing_ones, trailing_zeros,
unsafe_trunc, broadcast!_function, broadcast_function, broadcast_getindex, broadcast_setindex!,
cumsum_kbn, promote_shape, sum_kbn, blas_set_num_threads, escape_string, is_assigned_char, normalize_string,
print_escaped, print_joined, print_shortest, print_unescaped, print_with_color, unescape_string,
get_bigfloat_precision, set_bigfloat_precision, with_bigfloat_precision, get_rounding, set_rounding, with_rounding,
get_zero_subnormals, set_zero_subnormals,
plan_bfft!, plan_bfft, plan_brfft, plan_dct!, plan_dct, plan_fft!, plan_fft, plan_idct!, plan_idct, plan_ifft!,
plan_ifft, plan_irfft, plan_rfft, object_id, current_task, task_local_storage, time_ns, catch_backtrace,
promote_rule, promote_type, current_module, code_typed, code_warntype, code_lowered, code_llvm, code_native, module_name,
module_parent, __precompile__, include_string, include_dependency, gc_enable, nb_available, poll_fd, poll_file,
redirect_stderr, redirect_stdin, redirect_stdout, takebuf_array, takebuf_string, watch_file, init_worker, remotecall_fetch,
remotecall_wait, process_exited, process_running, disable_sigint, pointer_from_objref, pointer_to_array, pointer_to_string,
reenable_sigint, unsafe_copy!, unsafe_load, unsafe_pointer_to_objref, unsafe_store!

113 functions that are exported, many many more that are not exported (but since there is no way of telling whether they are "private" or not, they very well may be used outside of Base), and that also didn't count macros with _.

[dpsanders]

My takeaway from this discussion is:

• no ! on setprecision
• divided on underscores (I presonally tend to find readability increases with relevant underscores)
• replace with_precision with setprecision
• do the same for with_rounding and set_rounding.

I'll implement these in the PR, with no underscores for now.

[StefanKarpinski]

@ScottPJones: we would like to get rid of or rename all of those. Underscores are a sign of an old, crappy or inadequately designed API. By Julia 1.0 we should have eliminated most of them. In particular, we should not be adding more of them.

[simonbyrne]

Okay, I'm fine with this. Would you mind rebasing into one commit?