On Mon, Sep 13, 2021 at 11:32:56AM +0200, Pavel Machek wrote:
Hi!
quoted
quoted
quoted
config WERROR
bool "Compile the kernel with warnings as errors"
- default y
+ default COMPILE_TEST
That seems reasonable. It very much is about build-testing.
That and 2 more things IMO:
a. having developers be responsible for build warnings, not just
build errors
b. having maintainers merge them more like they are build errors
and not just some warnings that can be overlooked.
I don't see enough of a. or b. :(
Do we really want developers treat warnings as errors? When the code
is okay but some random version of gcc dislikes it...
Plus, there's question of stable. We already get ton of churn there
("this fixes random warning"). WERROR will only encourage that...
I will not be backporting this patch to older stable kernels, but I
_want_ to see stable builds build with no warnings. When we add
warnings, they are almost always things we need to fix up properly.
Over time, I have worked to reduce the number of build warnings in older
stable kernels. For newer versions of gcc, sometimes that is
impossible, but we are close...
thanks,
greg k-h