Sympathy with killers can appear in artistic or critical contexts in which the relation to the overall standpoint of the author is subtle and complex. Suppose a poem or a novel is written from the point of view of someone sympathetic to, say, Palestinian suicide bombers. How explicit does the detachment between the work and the author's real-life view for this to count as legal? Furthermore, advocacy of violence is clearly not banned in general in France. This is a country with a major tradition of intellectuals supporting violent regimes, and in which tactics in wars are debated openly. I guess the response will be that in these cases, the deaths of civilians are mere by-products of tactics rather than the end goal as with Charlie Hebdo. But really it is not so clear that the Charlie Hebdo killings are not tactics in a larger geopolitical effort. Prosecutors and bureaucrats are ill-placed to adjudicate these boundaries in binary fashion, and the risk that the law will be applied asymmetrically from issue to issue and victim to victim is serious.
Of course, the existence of difficult borderline cases does not automatically make for bad law. But on the converse side, it is pretty unclear what the positive effect of these hate laws is. When one compares the situation of Islamic fundamentalist radicalism or the racist far right in Europe and the US, it is hard to reach the conclusion that these laws are actually lessening the attitudes they target. Granted, there are many other social variables. But other areas of contrast between American permissiveness and the restrictiveness of other democracies - say on guns or campaign finance (no First Amendment absolutist am I, on this issue!) - there are clear distinctions visible in the actual effects on gun violence or money in politics. These hate laws appear much more ineffective.
It is not just support of violence, but support of terrorists attacking your own country. That is treason.