Code review comment for lp://staging/~osomon/oxide/maxCacheSize

I think the current approach is ok - it's preferable for argument validation to go near to their API's in qt/ because we can be consistent with the logging mechanism we use (although, this could probably actually go in WebContextAdapter). I'd be tempted to just DCHECK that the passed in size isn't too big in BrowserContextIOData though, to make it clear to anyone who wanted to write a new port that they need to limit the size.