Unflushed Chunks

[kmdavis]

Sometimes dust gets stuck in an asynchronous block at the end of a template. We have our own subtemplate/partial include system, and sometimes, for some templates, after ending a stub, it doesn't get flushed.

I fixed the issue locally like this, but it's kinda scary: (works for our use cases though)

Stub.prototype.flush = function() {
  var chunk = this.head;

  while (chunk) {
    if (chunk.flushable) {
      this.out += chunk.data;
    } else if (chunk.error) {
      this.callback(chunk.error);
      this.flush = function() {};
      return;
    } else {
      // Sometimes, dust get's stuck here, at the end of a template, and something thinks there's more to do
      // chunk.next is undefined
      if (!chunk.next && "" === chunk.data) {
        break;
      }
      return;
    }
    chunk = chunk.next;
    this.head = chunk;
  }
  this.callback(null, this.out);
}

Also, for reference, here's our custom include helper:

dust.helpers.include = function (chunk, context, bodies, params) {
  return chunk.map(function(chunk) {
    var ctx = (params && params.context) || context.get("context") || context.current();
    ctx.currentDustIndex = context.stack.index;
    ctx.currentDustLength = context.stack.of;
    Gilt.Template.render((params && params.name) || context.get("name"), ctx, function (html) {
      chunk.end(html);
    }, {
      engine:  (params && params.engine)  || context.get("engine") || "dust",
      version: (params && params.version) || context.get("version")
    });
  });
};

And an example of it in use: {@include name=\"cart\" version=\"countdown\" context=sku/}

[kmdavis]

Actually, this hack doesn't work for our usage. It produces output, but the output isn't in Quite the right order

[kmdavis]

Ok... fixed it locally. Reverted that hack above. In my usage, I had an if statement wrapped around the include. By moving the if statement into the partial, it works. But it would be nice not to have to do so