Developing a device driver for BSP

[Gedare Bloom]

Hi Steve,

You can usually get away with 'make' in your build directory, which
should only compile dependent files, followed by 'make clean' and
'make' in the testsuites if you are using them, followed by 'make
install' if you want to install the new version to your local system.
Then you could re-link an application against the installed RTEMS if
that is how you are testing. (I usually stick application code into
testsuites so I can build it with RTEMS while developing.) The major
weakness of the current build system for this kind of development loop
is that the dependencies for changed files do not get picked up in the
testsuites / linking of executables. Thus the 'make clean' in
testsuites if you make a change in the other parts of RTEMS.

I often develop and test directly from a build directory. If you only
have one 'configure' command line, the only times you need to re-run
configure is if you change autotool inputs (Makefile.am, configure.ac
files), which should also be accompanied by a bootstrap -c, bootstrap
-p, and bootstrap. (You can also run these from any subdirectory in
RTEMS that contains a configure.ac to only update due to changes under
that subdirectory's autotool sources.)

Gedare

On Mon, Dec 15, 2014 at 4:23 PM, Steve B <sbattazzo at gmail.com> wrote:
> Hello all,
>
> Quick question regarding the development flow for a device driver to go with
> a BSP.
> Would I need to put the source in the libbsp/(arch)/(bsp)/ and rebuild the
> BSP every time something changes, or could I develop the thing as a sort of
> standalone application and then move it into the BSP tree when it's
> finished?
>
> I might like it if I could do the latter because building the BSP seems to
> take quite a bit of time, unless the make process can see that only
> something small has changed and only rebuilds what it needs to. I haven't
> tried it yet so I'm not sure what's going to happen.
>
> Cheers,
>
> Steve
>
> _______________________________________________
> users mailing list
> users at rtems.org
> http://lists.rtems.org/mailman/listinfo/users