Alias String#gsub to replace or deprecate gsub

[pynixwang]

it 's best time to do this deprecation before 1.0

ref:
https://en.wikipedia.org/wiki/Comparison_of_programming_languages_(string_functions)#replace
https://forum.crystal-lang.org/t/alias-string-gsub-to-replace-or-deprecate-gsub/2538
#8452

[pynixwang]

some people do not like deprecation, so alias.
like gsub use gsub
replace for beginners. and other language people.

[NIFR91]

I vote against alias, I've always agree with crystal's non/mininum-alias policy, it is great to only have one way, easier for beginners and for experienced crystal programmers.

It should be either left as sub and gsub or change (deprecated) to the new name (replace and greplace?).

[straight-shoota]

Adding an alias is not productive and we are commited to avoid aliases in Crystal. So I don't think it's worth considering this option.

Renaming to replace might have some benefit of being easier to understand because it's common in other languages.
But "(g)sub" is short for "substitute" which describes the same process as "replace" in natural language. Since it's shorter, it's less to type.

In general I'd still agree that "replace" might be the better choice than "sub". If we were at the drawing board of a completely new standard library, I would vote to use that.
Yet overall the advantages and disadvantages are not very strong for either side. But there's a strong reason to keep the status quo: avoid migration costs.

These methods are very commonly used and renaming them would introduce breaking changes in a large number of code bases.
I don't think the potential benefits can outweigh the associated migration cost by far.

[pynixwang]

migration cost is lowest before 1.0

[rubyFeedback]

we are commited to avoid aliases in Crystal.

https://crystal-lang.org/reference/1.5/syntax_and_semantics/alias.html

alias PInt32 = Pointer(Int32)

I don't really mind that you guys went the "we don't use aliases" - but
if even the docs mention that you DO use aliases, then it's a bit
weird and inconsistent to try to "reason" that "we don't use aliases"
when ... uhm ... you actually do.

To me personally it seems that you do use aliases, but just not for
methods but for types.