For those of you who have been struggling to get the ESPRESSIF AT command set working with the pesky BUSY…S problem on the little ESP8266 and similar boards, I’ve been testing various versions with them and last week they produced what was called TEST3. The problem did not disappear entirely but it definitely stopped the ESP-01 boards from locking up completely – i.e. the busy message went away pretty quickly. I’m guessing you could poll it. In the AT demo my code worked. This morning they wrote to me to acknowledge that the current release of their software still has the BUSY…S issue but that this Friday they’ll be releasing a new version based on the test code. Espressif do of course also produce the SDK but you must have to be very patient to want to wade through all of that.
Suggest you keep an eye out for it.
Meanwhile the LUA implementation of the socket listener I detailed earlier is still working – I’ve left it sitting there and occasionally check my APP to see if it will talk to the board. It will. I have modified the code to actually turn 2 LEDS on and off – ie on ports GPIO0 and GPIO2 and they absolutely work. Elsewhere, Dave Allan has detailed how the soldering expert can grab back some more I/O pins. Use of the socket code can still lead to eventual memory degradation at this point however. I’m also having problems getting input from the serial line without it being interpreted.. so for example io.read() does not work – the interpreter seems to know nothing about io. Any input to the serial is run through the interpreter. That makes it pretty much impossible to use the code to talk to, say an arduino.
My Tech Weblog 
Holy Cow, could it be…. Friday can’t come soon enough!!! I hope to God that this is a real fix. In the mean time, where can one get the latest test code to play with???
Really appreciate your blog on this little bugger!
I’ve sent you an email… 🙂
Hello. I have been lurking here as well for the last couple weeks and watching for progress. I am most interested in the LUA project as I’m not sure I’m smart enough to deal with the other code. I currently have a “home brew” Arduino Pro Mini based automation system that communicates via 433Mhz modules – using VirtualWire. This of course requires a “hub” (Raspberry Pi) to accept a restful api request from a tablet or what have you, convert it to a string, and blast it out for the modules to ignore/respond as appropriate. To make a VERY long story short, I would love to convert everything to this module, keeping the Arduino but removing the need for a dedicated hub to translate messages from GET/POST data to 433Mhz strings. In all of this, how close do you think ware are to being able to GET/POST anything back and forth directly to a fixed IP address? By the way, based on your daily updates, I have purchased several of the modules in various flavors along with the little host boards, but I have set them aside waiting for viable firmware to emerge. Thanks for your willingness to be the pioneer on this and share your findings.
Isn’t that interesting – I have almost the same thing – but with 2.4ghz modules using RadioHead – and I was in the process of porting that to the 4333Mhz modules – and I too am thinking.. wouldn;t it be better to do the whole thing with WIFI modules – given their extreme low cost – I plan to make a library for Arduino for them to talk to each other. I didn’t need a Raspberry Pi for master as my “Arduino” home brews use 1284 chips which have bags of room and lots of pins. I think we’re not far off. Maybe Friday’s post might help. Also the fellow doing the LUA implementation might make a hit of that (I’m counting on it) so some simple high level code in those modules might help. There is currently a gaping hole in that software unless I’m missing something in that you can’t pass serial input straight through to the TCP/IP as it gets interpreted.. but I might be missing something or it may be he can graft something on. Either way I’m counting on miracles in the next week or two..