Detection fails on svg but don't know why about is-svg HOT 3 CLOSED

The svg is valid, according to W3C standards. The issue has to do with the regex being used:

/^\s*(?:<\?xml[^>]*>\s*)?(?:<!doctype svg[^>]*\s*(?:\[?(?:\s*<![^>]*>\s*)*\]?)*[^>]*>\s*)?(?:<svg[^>]*>[^]*<\/svg>|<svg[^/>]*\/\s*>)\s*$/i

As it is, due to this part of the expression: \[?(?:\s*<![^>]*>\s*)*\] it does not allow for markup to be included in an Entity's internal subset. If you remove them, isSvg() returns true.

To fix this we could simply remove > has a forbidden character to occur inside the square brackets of the doctype tag (with the end expression being: /^\s*(?:<\?xml[^>]*>\s*)?(?:<!doctype svg[^>]*\s*(?:\[?(?:\s*<![^]*>\s*)*\]?)*[^>]*>\s*)?(?:<svg[^>]*>[^]*<\/svg>|<svg[^/>]*\/\s*>)\s*$/i

Or update the regex to allow for markup. I tried to come up with something. but allowing for markup would mean we would need to allow for pretty much anything. So I went with allowing any content to be added when inside double quotes. Like so:
/^\s*(?:<\?xml[^>]*>\s*)?(?:<!doctype svg[^>]*\s*(?:\[?(?:\s*<!\s*[^]*(?:\"[^]*\")*\s*>\s*)*\]?)*[^>]*>\s*)?(?:<svg[^>]*>[^]*<\/svg>|<svg[^/>]*\/\s*>)\s*$/i

But this would make the function even dirtier as it would allow for any type of scrap to happen inside the doctype tag. Are any of these solutions viable? Any ideas? I can open a PR and address this.

from is-svg.

@eluisfonseca How about a regex that strips out markup (or just everything) from inside the entity before running the detection?

from is-svg.

Good idea. So we'll do no check of the entity's content and just remove it and then just validate the what is left. I'll look into it.

from is-svg.