Hi Graeme,
On Wednesday, April 20, 2011, Graeme Russ graeme.russ@gmail.com wrote:
On Wednesday, April 20, 2011, Detlev Zundel dzu@denx.de wrote:
Hi,
As a base for discussion, what about this:
Use common sense in interpreting the results of checkpatch. Warnings that clearly only make sense in the Linux kernel can be ignored. Also warnings produced for _context lines_ rather than actual changes can also be ignored.
One man's common sense is another's idiocy
I vote for a zero warnings, zero errors U-Boot specific checkpatch
I also think that all patches should be submitted with a checkpatch summary with an explaination for any errors or warnings - this will at least save a little effort for the maintainers and reduce the number of patches bounced only to have the checkpatch problems argued away by the author anyway
When we accept 0 errors and 0 warnings only, then we will always see the same text :)
As long as we are not there, I do agree but then we should come up with a recipe on how to automate this. I looked into git format-patch but it does not seem to have such an option. Does anyone have a clever one-liner for this?
Cheers Detlev