On Tue, May 13, 2014 at 6:00 AM, Anne van Kesteren <ann...@annevk.nl> wrote: > On Tue, May 13, 2014 at 10:33 AM, Boris Zbarsky <bzbar...@mit.edu> wrote: >> It can't work, given how URL.origin is currently defined... It's possible >> that definition should change, though. > > We don't want new URL() to take ownership of the Blob object, so > making new URL() reflect the origin of whoever created the uuid for > the Blob object seems weird.
Why would it need to "take ownership of the Blob object"? First of all, the origin of the blob: url is determined by who called createObjectURL, not who owns the Blob instance. Hence you don't actually need to touch the Blob instance to figure out the origin, but rather just inspect the url itself (if we use explicit origin syntax), or look up the origin in the internal url->Blob table (if we use implicit origin syntax). And even if you did somehow need to touch the Blob, the implementation could just immediately release it before returning from the constructor. / Jonas
