order of \not character reversed on unicodeit.net

[rigille]

Typing

0 \not\in \mathbb{N}

I expected to get

0 ∉ ℕ

Instead I got

0 ̸∈ ℕ

and

0 \in\not \mathbb{N}

yields the expected result. So maybe the order of the \not character should be swapped with the next character?. By the way 0 ∈ ℕ hahha

[cemulate]

Just leaving some info after my own investigation:

\not\subset

produces

\xCC\xB8\xE2\x8A\x82

Which renders incorrectly in most contexts as "̸⊂"

\xCC\xB8 is a "combining long solidus overlay" and the rest is "subset of". According to the standard, combining characters should come after the original. So though this is a faithful literal rendering of the input, it is not the "correct" or expected result.

This is why if you use

\subset\not

The literally-rendered result will be

\xE2\x8A\x82\xCC\xB8

Which contains the combining character after the subset character and produces the correct result: "⊄"

---

I see two options for a "fix":

• Catch situations in which a TeX code will render as a combining character and appropriately switch the order
• Catch common commands like \not\subset and render them as the existing unicode character for that combination, this could be rendered as the individual "not a subset of" character which has its own codepoint, instead of using combining characters.

[cemulate]

I submitted a PR #61 to handle cases like this.

[HDembinski]

This should be closed since the PR was merged.