NetBurner Community Forum

Hi, I'm glad to say my current project reached a milestone today, and will get more complicated soon. I think its a good time to consider version control, with perhaps a change of host operating system for maturity purposes. Could I ask the netburner community for general knowhow?

1. I'm a Unix guru, and can do a lot with command line utilities in generating scripts and so forth. Is there a workflow to generate a Unix toolchain to generate motorola code with the NNDK uC/OS libraries (Cross-compiling ?)

1a. If I have a mixed host environment, and both Windows, Unix toolchains are possible, are there any pitfalls?

2. I saw CVS support (haven't used it) in the NNDK Eclipse. I think for a modern project, with distributed team members, Git might be suitable for my future application development workflow:

2a. Is there a supported way to incorporate the checked-out git version info into the NNDK generated makefile so the banners of each code module can display a consistent version number ?

2b. I tried to incorporate Egit into the NNDK Eclipse system by utilizing (possibly incorrectly ...) http://www.banym.de/eclipse/install-git ... or-eclipse to load up http://www.eclipse.org/egit/ . Is this permitted with the NNDK as distributed by Netburner? I have an older machine at home: 2.5.3. Is this possible with a newer NNDK version?

2c. If (1a) and (2b) not possible, can the functionality of the Netburner Eclipse IDE toolchain/workflow be replicated with a stand-alone installation of standard Eclipse + Egit + GCC + xxxx ?

3. Finally (off-topic) if I had to use Java (for the client side apps) and NNDK C++ development (for the embedded server apps) toolchains using an Eclipse-like environment, would you recommend separate eclipse installations for the two universes, or one ? I'd like one, to have consistency between the two platforms, as the two applications (client -- server) should IMHO be developed in parallel. But, possible ?

Thank you for your time. I apologize if my questions are newbie like, but I can only code well when I have a framework to code in. Any tips appreciated.

You'll probably want to take a glance at the wiki article on using standard Eclipse releases. I have found using NBEclipse with additional plug-ins problematic (but I haven't tested it recently). My suggestion is to stick with the CDT Indigo release of Eclipse. I think Eclipse EGit comes with the standard CDT Indigo release. I prefer HG so I use MercurialEclipse.

I use pre-build steps to incorporate HG version info into my overall version information. I have commands that report the firmware version to the user. My philosophy is to not litter my source code with version information, to me that is one reason why I use a source code control system, and that is where the revision information belongs. You can tell MercurialExlipse to show the version info automatically in the project explorer. By default it only shows the latest revision of the project but a simple checkmark will show the current revision for each file. See the image. The overall project is at 44 but some files shown are at revisions: 14, 17, 42 and 44.

Regarding item 3. I would probably use one installation but two workspaces.

tod wrote:You'll probably want to take a glance at the wiki article on using standard Eclipse releases.

Looks very interesting, thanks!