I figure this will probably be fixed in the 1.2 release next month, but wanted to ask whether there was a workaround.
I implemented a socket client, which connects to a server and sends commands and responds to commands.
When it first connects, it sends two commands to the server. The onWrite event only fires when it reads something coming in from the server. Then the server sends something I have to respond to, so another send. This time there isn't anything coming from the server, so instead of writing when it finally does get something it just times out.
Pretty sure this is a bug in the API because my code doesn't do anything the tests in drillbit don't do besides expect to be able to do a send when nothing is coming down the pipe. The way the onWrite fires at exactly the same time as the onRead and fails to fire when there is no onRead (because no data coming from server at the time) seems to indicate they are dependent on each other when they shouldn't be.
I can post the code (it's actually a port of already OSS software) but it really isn't doing anything tricky, just normal socket client functionality.
Unfortunately the 1.2 update was mobile only not both mobile and desktop like I thought it would be… don't know why I thought desktop was included.
I want to add on to the desktop socket test to make it test more than just send and get something back because the problem I'm having (which I'm sure it will also share once I add the test) is not with that initial case (my implementation does send and get something back, it is subsequent sends that time out) so I think it is a case of the test and the way most people are using it (page from web server) that is incomplete to fully test the socket functionality.
Will update JISIRC with the port and link to it here one of these days, maybe it'll turn out to work for others and just be a problem with my setup.