TASTEME deobfuscate, comment, and add non-ascii behavior

[roleohibachi]

An obfuscated meme function called TASTEME() was introduced ~5mo ago with no documentation. It appears to check for 3 sequential characters in a string; however its behavior is irregular for non-ascii strings, and could be dangerous depending on user implementation. For example, if used for hyperlink validation, it would be vulnerable to IDN homograph attacks.

For example:

input	actual result	expected result
abcddd	True	True
aabbcc	False	False
https://www	True	True
://	None	False
abcçççddd	None	True
100700070006c0065	None	True

This PR de-obfuscates the code, and uses enumerate() and the == operator to make comparisons. This results in more sane/safe behavior.

Since this function was introduced as a meme, it should probably just be removed. But that's like, just my opinion, man.

[paulfitz]

Thanks @roleohibachi ! A lot to unpack here. A small technical issue - I'm seeing strings like Eye of the tiger and Survivor classified as true by TASTEME? https://gristlabs.github.io/grist-static/weresheet.html?p=2#a1.s3.r19.c9

[alexmojaki]

It appears to check for 3 sequential characters in a string

You're close, but I don't think that's correct. I think it's quite fun puzzling out what it's doing, I just enjoyed the experience a second time. 👎 on deobfuscating.

On the other hand, 👍 for overengineering a pointless meme. I'd be in favour of a change that keeps to the style of the original code while handling non-ascii characters better, e.g. removing accents the way that _sanitize_ident does.

[roleohibachi]

"Won't fix". I will spend my time on other parts of the project.