Once I have a HTTP GET request block in the planner setup I can’t change the URL anymore. I then have to remove the block and add an empty new one. That then changes the ‘target’ variable coming in on the receiving web server. I would like to be able to change the URL without delete/add. And furthermore I would like to receive some ID related to the imp wifi card thing and not some object in the planner.
Finally why is it not possible to GET directly from the imp or at least the code. Now I have this nice coding environment and then I have to start dragging stuff with the mouse. Also the service I need to GET to is on a local network, now I have to route things in in a weird way. At some point they even blocked electric imp’s IP because to much GET’s where coming from one IP. So I had to change sites to keep on going.
I can change the URL in the HTTP Request block. The URL for the HTTP In block seems to be generated when you place the block. As far as i can tell there is no need to change this URL. They seem to be using random and unique numbers in the URL to make it safer.
It’s not the imp that does the HTTP GET, that is done in the planner. The imp only talks to the imp support infrastructure. That way the solution you build with an imp only need wifi access and does not need to know anything else about the local stuff.
Some posts mention the HTTP stuff will be improved in a newer version and i have also seen comments on planner side scripts called agents. The idea seems to be that the imp does the local IO and processing and the agents do the cloud side processing.
The imp is not an arduino, Raspberry Pi, mbed or beagleboad. It seems to be intended to connect physical local devices to services in internet. I like the idea that it cannot connect to devices on my local network and intend to set up an OpenWRT wireless router to keep the imp on its own network.
Yeah but that is a shame as none of the other devices are so easy in setup. In this case just some blinking screen and GO! And also the whole website cloud thing to write your software is genius as you don’t have to setup GNU toolchains, figure out JTAG programmers and so on.
But how hard would it be to allow the imp to directly GET to some local website. This way a crappy uplink (and I know many) do not brake my system. When I want to switch on the light the message now goes all the way to California and back.
So the whole development and downloading firmware to the imp can stay the same, but in my case once deployed it should not matter that the internet connection drops. I am convinced there are more setups like this.
BTW. You said: I can change the URL in the HTTP Request block. But can you also once it did a valid GET and got a 200?
Whilst it wouldn’t be hard to make the imp do local TCP stuff, that’s not what the imp does, as that’s not applicable for the majority of end users; we’re concentrating on the zero-setup market and believe we’ve got something that’s pretty awesome there.
In the future there are likely to be some interesting developments in the local reliability space but I can’t say more than that right now.
On your last question: I’m not sure what you’re asking?
I add an HTTP request block to the planner, type in a valid URL and set it to GET. Then use it and it works (get 200 in the box). When I then want to change the URL it won’t save the new url. It keeps on showing and using the old URL. Meaning I have to remove the block and add another one when I want to switch URLs.
Ah, I see. There’s a lot of things that are a bit screwy about the current front-end. You may have luck if you hit enter after typing the url, then clicking within the popup to drive some events that’ll cause it to get saved.