added url parsing to allow query params in urls

[digitalLumberjack]

Closes #434

Url is now parsed and query params are removed when checking for file extension.

The remaining issue is if someone would use query params to specify the file like htttps://mysditribution.com/download?file=root.tar.xz

What do you think ?

[XECDesign]

The change looks good to me. Has it been tested?

[digitalLumberjack]

Just thought about something, should we also support that for marketing, os_info and partitions files ?

[digitalLumberjack]

For now, I tested with a os_v3.json like this one :

{
    "os_list": [
	{
            "description": "Recalbox, the open source retro gaming console for rpi1!",
            "icon": "https://install.recalbox.com/v1/noobs/recalboxOS.png",
            "marketing_info": "https://install.recalbox.com/v1/noobs/marketing.tar",
            "nominal_size": 2048,
            "os_info": "https://install.recalbox.com/v1/noobs/rpi1/os.json",
            "os_name": "recalboxOS-rpi1",
            "partition_setup": "https://install.recalbox.com/v1/noobs/partition_setup.sh",
            "partitions_info": "https://install.recalbox.com/v1/noobs/partitions.json",
            "release_date": "2016-11-26",
            "supported_hex_revisions": "2,3,4,5,6,7,8,9,d,e,f,10,11,12,14,19,0092,0093",
            "supported_models": [
                "Pi Model",
                "Pi Compute Module",
                "Pi Zero"
            ],
            "tarballs": [
                "https://install.recalbox.com/v1/noobs/rpi1/boot.tar.xz?source=noobs",
                "https://install.recalbox.com/v1/noobs/rpi1/root.tar.xz?source=noobs"
            ]
        }
    ]
}

[XECDesign]

It would make sense to be consistent, but I don't have any strong feelings one way or another.