Registry download count number accuracy

[dangoor]

Some of the download count numbers seem inflated. Specifically, the numbers since May seem too high. Here's Markdown Preview:

        }, {
            "version": "1.0.1",
            "published": "2014-03-29T16:27:25.497Z",
            "brackets": ">=0.37.0",
            "downloads": 1839
        }, {
            "version": "1.0.2",
            "published": "2014-04-08T23:03:48.986Z",
            "brackets": ">=0.37.0",
            "downloads": 3204
        }, {
            "version": "1.0.3",
            "published": "2014-05-07T16:58:33.321Z",
            "brackets": ">=0.37.0",
            "downloads": 282737
        }, {
            "version": "1.0.4",
            "published": "2014-05-30T13:45:20.333Z",
            "brackets": ">=0.37.0",
            "downloads": 169369
        }, {
            "version": "1.0.5",
            "published": "2014-06-03T21:08:49.714Z",
            "brackets": ">=0.37.0",
            "downloads": 205019
        }, {
            "version": "1.0.6",
            "published": "2014-06-09T00:27:22.542Z",
            "brackets": ">=0.37.0",
            "downloads": 295730
        }],

Notice that 1.0.2 logged 3,200 downloads in a month. 1.0.3 logged 282,000 in just over 3 weeks. Either there was a massive spike in the number of people using Markdown Preview, or the numbers are off.

cc @ingorichter

[peterflynn]

Could it be some sort of bot that hit the S3 download bucket? Certainly seems to be a ubiquitous problem these days...

[ingorichter]

This PR adobe/brackets-registry#56 will fix the issue.

[ingorichter]

This will fix this issue adobe/brackets-registry#58