This is partially related to #50 -- the patches to get irmin-test in the browser will take some time to upstream probably (if at all desired by the the irmin folk), but I did manage to get it running to test the websocket backend in the browser. The main problem was the growing of the buffer, which was wrong.
I also took the liberty of catching any errors in the webwocket creation in the browser, previously we would just end up stuck on a promise that never resolves. At the moment, the irmin-test suite doesn't pass in Chrome but does in Safari. Chrome is closing sockets with code 1006 -- I think this might be some kind of security feature because we open lots of connections which is a little annoying.
This is partially related to #50 -- the patches to get
irmin-testin the browser will take some time to upstream probably (if at all desired by the the irmin folk), but I did manage to get it running to test the websocket backend in the browser. The main problem was the growing of the buffer, which was wrong.I also took the liberty of catching any errors in the webwocket creation in the browser, previously we would just end up stuck on a promise that never resolves. At the moment, the
irmin-testsuite doesn't pass in Chrome but does in Safari. Chrome is closing sockets with code1006-- I think this might be some kind of security feature because we open lots of connections which is a little annoying.cc @dinakajoy