Did the discussion about redundant parameter validations and DEVELHELP die?
I like the idea of getting rid of some redundant input validation. For example, if you are internally using spi_transfer_byte to provide spi_transfer_regs, then if the SPI device is valid for the first byte transferred, then it is probably going to be valid for the rest of the bytes in the same function call chain. There was some discussion about null pointer checks in a PR or a mailing list thread but I did not find it when I did a brief search. Best regards, Joakim Gebart www.eistec.se On Wed, Mar 25, 2015 at 5:02 PM, Kaspar Schleiser <kas...@schleiser.de> wrote: > Hey, > > On 03/25/2015 11:12 AM, Hauke Petersen wrote: >> >> in general I like the idea, one problem I see is however, that is not >> always clear, to which level an API belongs (e.g. the GPIO API is >> definitely used also by high-level application programmers, while still >> belonging to the low-level peripheral drivers...). > > We could mark certain functions / parts of an API as "advanced" in the docs > and provide "safe" alternatives. > > Seriously, it hurts to not be able to work around 1000000000 redundant > checks whether an int coming from flash is a correct SPI device... > > > Kaspar > _______________________________________________ > devel mailing list > devel@riot-os.org > https://lists.riot-os.org/mailman/listinfo/devel _______________________________________________ devel mailing list devel@riot-os.org https://lists.riot-os.org/mailman/listinfo/devel
