Suggestion: Ability to use arrays of child-nodes everywhere a single node can be used.

One suggestion: It would be helpful if vnodes could insert arrays of vnodes instead of only a single one. This would make it much more flexible and would eliminate the need for “useless” in-between container-tags (which genrally works fine, but is problematic for certain use-cases for example when using a nth-child-selector in css).

To circumvent this, in the moment I use a simple ‘h’-wrapper which flattens it’s arguments before inserting:

function flatten(data) {
    var res = [];
    function flat(data) {
        _.each(data, function(v) {
            if (v === null || v === undefined || v === false) {}
            else if (_.isArray(v)) flat(v);
            else res.push(v);
        })
    }
    flat(data);
    return res;
}
exports.h = function(tag, opts, args) {
    return h(tag, opts || {}, flatten(Array.prototype.slice.call(arguments, 2)))
};

This makes building node-trees much easier (but requires a {} if there are any child nodes).

Problem with my solution: It won’t work for thunks, because data.vnode only supports a child-node. So if there is a new thunk-implementation comming, maybe it’s possible to address this problem, too.

[paldepind]

Many utility libraries, both Ramda and Lodash, have a flatten function. I.e. one can just do h('div', flatten([...])). This seems like a reasonable solution to me so I’ll close this. Feel free to reopen if you disagree 😄

[TylorS]

@kay999 vnode is now part of the documentation.