On 18/06/18 05:20, Luc Van Oostenryck wrote: [snip] >> Hmm, I think "different restricted types" or "different restricted >> base types" would work for all messages. > > There is two different cases: > 1) '__be32' vs. 'int' > 2) '__be32' vs. '__le32' Indeed. > In 1), __be32 is restricted while int is not. In 2) both types are > restricted but are different restricted types. I would have a distinct > error message for them. I guess I don't see a great need for different messages. > BTW, I'm not at all found of the use of the word 'restricted' in the > error message (and sparse's code) while the corresponding attribute > is 'bitwise' and 'restrict' is an unrelated keyword since C99. Yes, that caused me to stumble quite often in the first few years! :) So, I agree it would be good to fix up the code ... > So maybe we should use instead "different bitwise types" and ... > "different bitwiseness"? ;) ... and my only concern with new wording in the messages - would this cause kernel developers to double take? (non-kernel developers would most likely not be affected, since they don't use 'bitwise' types anyway). Having said that, I much prefer something like: 1) "mixing 'bitwise' and normal types" 2) "different 'bitwise' types" Or, something like that. ;-) ATB, Ramsay Jones -- To unsubscribe from this list: send the line "unsubscribe linux-sparse" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
