[Buildroot] svn commit: trunk/buildroot/scripts: test
ulf.samuelsson at atmel.com
Sat Jan 24 16:45:26 UTC 2009
lör 2009-01-24 klockan 12:50 +0100 skrev Peter Korsgaard:
> >>>>> "ulf" == ulf <ulf at uclibc.org> writes:
> ulf> Author: ulf
> ulf> Date: 2009-01-24 11:29:10 +0000 (Sat, 24 Jan 2009)
> ulf> New Revision: 25005
> ulf> Log:
> ulf> Add test configuration but do not include it by default
> ulf> Added:
> ulf> trunk/buildroot/scripts/test/
> ulf> trunk/buildroot/scripts/test/Config.in.test.buildall
> Come on, just keep it locally instead if it isn't used by BR.
You can say the same for all the defconfigs.
Buildroot does not use them unless you do a make *_defconfig
In the same way, a BR user will be able to do a test
of a specific board/architecture by running dedicated commands.
> What does this buy you compared to just having a .config with those
> things enabled?
It saves editing configs.
I can apply this to ANY board, without manual editing,
and there are a lot of boards/architectures.
By using several Config.in.test files, it is easy to
test a limited set of things for many architectures.
> Please discuss these things on the list first when you know they are
I already suggested this on the list and got no response.
What is the benefit of NOT having selftest included?
I have pointed out that so far we have no documentation
on what works and what does not work.
If the results of running these scripts are checked in
then we *have* documentation which is generated automatically.
Do you have ANY way of measuring the quality of Buildroot
before you do a release, or do you just plan to
tell everyone that we are creating a release
and then take the trunk and publish it a certain date
regardless of quality?
More information about the buildroot