Bug: parsing negative attributes in tag definition

The code:

el("input[type=number][min=-1]");

Generates:

<input type="number"></input>

Instead of:

<input type="number" min="-1"></input>

Example: https://domvm.github.io/domvm/demos/playground/#mobx

[jlgrall]

The advantage with the simpler regexp is that it doesn’t skip attributes that contain unsupported characters. It took me a few days to notice that the [min=-1] was not rendered, because I forgot that I wrote it. It is always better to not ignore errors.

I would also accept a = without a value. After all, HTML accepts empty attributes, so it is less surprising to accept it too.

i would add \s to the attr name exclusions

If you accept: el("div", {"at tr": "val"}), I think you can also accept el("div[at tr=val]"). At least this way it has a chance to trigger an error on the client instead of being completely ignored.

Which would give us:

/\[([^=\]]+)(?:=([^\]]+)?)?\]/g

[jlgrall]

Looks good.