Fix split / rsplit collision

[CHB-0r1s]

After research of all possible combinations, there is such schema for prevent useless mutations:

Without mutations

1. split() -> split()
2. rsplit() -> rsplit()
3. split(" ") -> split(" ")
4. rsplit(" ") -> rsplit(" ")
5. split(sep=" ") -> split(sep=" ")
6. rsplit(sep=" ") -> rsplit(sep=" ")
7. split(maxsplit=-1) -> split(maxsplit=-1)
8. rsplit(maxsplit=-1) -> rsplit(maxsplit=-1)
9. split(" ", maxsplit=-1) -> split(" ", maxsplit=-1)
10. rsplit(" ", maxsplit=-1) -> rsplit(" ", maxsplit=-1)

With mutations

1. split(maxsplit=1) -> rsplit(maxsplit=1)
2. rsplit(maxsplit=1) -> split(maxsplit=1)
3. split(" ", 1) -> rsplit(" ", 1)
4. rsplit(" ", 1) -> split(" ", 1)
5. split(" ", maxsplit=1) -> rsplit(" ", maxsplit=1)
6. rsplit(" ", maxsplit=1) -> split(" ", maxsplit=1)

I save it as a note in comments.

[CHB-0r1s]

I hope this fix can help with #389

[CHB-0r1s]

@Otto-AA @boxed This magic with parsing -1 and +1 in libcst parser cause a lot of troubles for me. So, please, if you know better way to do it, I appreciate your help.

[Otto-AA]

Hi, thank you for this PR!

@Otto-AA @boxed This magic with parsing -1 and +1 in libcst parser cause a lot of troubles for me. So, please, if you know better way to do it, I appreciate your help.

I don't think that x.split(maxsplit=-1) is written directly in the code. This is equivalent to x.split(), which I would expect people to use, rather than explicitly providing and disabling maxsplit.

So I would suggest that we do not check if the value is -1 and simply always mutate, if maxsplit is provided with any value. So x.split(' ', -1) would also be mutated to x.rsplit(' ', -1), which is not optimal but should happen very seldomly.

[CHB-0r1s]

Thank you @Otto-AA !
I will apply your suggestions as soon as possible today)

[CHB-0r1s]

So, new schema

Without mutations

1. split() -> split()
2. rsplit() -> rsplit()
3. split(" ") -> split(" ")
4. rsplit(" ") -> rsplit(" ")
5. split(sep=" ") -> split(sep=" ")
6. rsplit(sep=" ") -> rsplit(sep=" ")

With mutations

1. split(" ", 1) -> rsplit(" ", 1)
2. rsplit(" ", 1) -> split(" ", 1)
3. split(maxsplit=1) -> rsplit(maxsplit=1)
4. rsplit(maxsplit=1) -> split(maxsplit=1)
5. split(maxsplit=-1) -> rsplit(maxsplit=-1)
6. rsplit(maxsplit=-1) -> split(maxsplit=-1)
7. split(" ", maxsplit=-1) -> rsplit(" ", maxsplit=-1)
8. rsplit(" ", maxsplit=-1) -> split(" ", maxsplit=-1)
9. split(" ", maxsplit=1) -> rsplit(" ", maxsplit=1)
10. rsplit(" ", maxsplit=1) -> split(" ", maxsplit=1)

[CHB-0r1s]

Hello! I applied all suggestions. Code looks pretty easy now. Schema was removed from comments (add link).
Can you check new updates, please :3

@Otto-AA @boxed

[CHB-0r1s]

Just little ping @Otto-AA

[Otto-AA]

Thanks for the updates, looks good!

And no need to ping, whenever I have time I will go through all unresolved conversations. It can also take a week or two if I'm busy with life, but I won't forget you when I find the time :)