And everything is slightly related.

So first question, why on the LPC81x is for example the sleep API missing? Is it simply a matter of lack of time/interest to add it? Because of course not all devices can do the same, but it is nice if as much as possible is common for devices, and the LPC81x can certainly do sleep.

One of the answers on the previous question will be: "You can make it and then do a pull request". Indeed, thats what I did, I made sleep and deepsleep modes for it. Deepsleep needs a bit of polishing, so once that is done I can make a pull request. However I made it in the online compiler. And lets face it, the github mbed library is very active, and to me it looks like pull requests are fast processed. Pull requests on the mbed website are pretty much ignored, probably because you need to manually add it to github. But for someone like me who does use the online compiler, is there an easy way to go from my mbed-src code in the online compiler to a github pull request?

Finally, sleep, and specifically the sleep API. Can more sleep modii be added to the API? Devices do not need to have them implemented, but it allows for more options. Specifically for the LPC81x, deep powerdown is a PITA, however normal powerdown isn't harder than deepsleep, and consumes significantly less power. Or would it then be preferred to just implement powerdown then as deepsleep? So you stick to two sleep states.

Hi Erik. I think you have joined the team now :)

So, what is the answer to this question? Are Github pull requests still the quickest way to get changes into the source tree?

Thx.

Jerry.

This is an old post :).

But yeah, pull requests here are ignored (they should really close the ability and just redirect you to github), github pull request are generally handled quite fast, although it does really differ between requests.

Personally I have the mbed github cloned to my PC and also working build environment (ie, correct private_settings.py). If I want to change 1-3 files I just do it online, copy paste it to my local mbed and create a pull request. Then I also check if the modifications seem correct and I send the request.

For larger changes I either copy paste and try if it builds, or I completely make it offline (for example when working on a new target).

Hi Erik, and thanks.

I'm working with both the NXP 43XX and the Freescale K64 series processors.

Is Github then the preferred way to get code into the main distribution?

Also how about middle-ware stuff like webserver/websocket stacks? There does not seem to be a route from github to mbed.org for this stuff (?)

Any thoughts welcomed.

Yes github is the preferred way for that. There are also some other libraries which are maintained via that github (such as USBDevice). Things like webservers you can just publish directly here on the site.