Mfcttrf

On many places I may read that syntax is about structure, and semantics is about meaning, and this makes sense. But, lets think of the canonical example

Colorless green ideas sleep furiously.

said to be syntactically correct but semantically meaningless.

We can analyse this sentence as

adjective + adjective + noun + verb + adverb

and yes, is syntacticaly correct. But what if we syntactically analyse it as

adjective + adjective-for-chromatic-things + abstract-noun + passive-verb-for-concrete-things + active-adverb

or something along those lines. We could divide the set of parts of speech into more specific ones, so that the sentence is syntactically only if it is semantically meaningful.

Is there a strict difference between the two? What could be the flaw in this example?

To answer the main question in the title: No, there is no strict wall between syntax and semantics. I think anyone who has worked at all in syntax will know that syntax, semantic and pragmatics are intricately linked, and you cannot study one without the rest. For example, consider the Manipuri case marker nǝ (Bhat, 2002) (this is entirely a random example that's only chosen because it's something I've been reading about and it's a clear-cut example):

(1) ma-nǝ ǝy-bu kawwi
 he-nom me-acc kicked
 ‘He kicked me’

(2) ma ǝybu uy
 he me-acc saw
 ‘He saw me’

The reason we add the marker in (1) and not (2) is because (1) is about an activity and (2) is about a state. It is obvious from these and other examples that there is no strict separation between syntax and semantics.

However, the example that you give isn't such a great example. In principle, linguists only put words into parts of speech because of morphosyntactic criteria, i.e. criteria that are realised in the form of language, not semantic criteria, criteria that a purely based on meaning. (Examples of morphosyntactic criteria would be whether a word can modify a word of another type, whether a word can be inflected in a certain way, etc.) So we would not advocate for a category 'adjective for chromatic things' in English unless there is a specific syntactic reason for it, say if English colour adjectives were the only ones that can occur postnominally (which of course isn't true). Ditto for abstract nouns, concrete things, etc. Passive is a syntactic notion and not a semantic one, so it doesn't help your case (and I'm not sure what 'active adverb' means - 'furious' doesn't seem to involve action to me, and I'm not sure how it's related to active in the sense of active vs passive either).

Bhat, D. S. (2002). Grammatical relations: the evidence against their necessity and universality. Routledge.