It seems to me that some extended development effort will be needed for Cppcheck-1.88.
As you might have seen I am trying to move checkers from "simplified" to "normal". I want that in Cppcheck-1.88 we will not have Tokenizer::simplifyTokens2() anymore. At least not for the Cppcheck checkers. If there are rules that require the "simplified" token list we could run that.
Not sure if we can do this now... but I would like to remove our old analysis for uninitalized variables. We have ValueFlow analysis in parallell, and I only want to rely on that. I think it will take some effort to fix false negatives. However these false negatives affect all our other checks today.
We currently have the --remove-unused-templates option. imho that looks promising but it needs to be verified that it works properly. I don't know if we want to have a separate option for this. Do you have any opinions?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I am working on a major rewrite of CheckBufferOverrun.. there is lots of old code that should have been rewritten long ago. Now when moving it to 'normal' it's difficult to avoid a rewrite.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
It seems to me that some extended development effort will be needed for Cppcheck-1.88.
As you might have seen I am trying to move checkers from "simplified" to "normal". I want that in Cppcheck-1.88 we will not have Tokenizer::simplifyTokens2() anymore. At least not for the Cppcheck checkers. If there are rules that require the "simplified" token list we could run that.
Not sure if we can do this now... but I would like to remove our old analysis for uninitalized variables. We have
ValueFlow
analysis in parallell, and I only want to rely on that. I think it will take some effort to fix false negatives. However these false negatives affect all our other checks today.We currently have the --remove-unused-templates option. imho that looks promising but it needs to be verified that it works properly. I don't know if we want to have a separate option for this. Do you have any opinions?
It would be good to still have Tokenizer::simplifyTokens2() for the rules since those cannot use ValueFlow and are limited to lexical matching.
I am working on a major rewrite of CheckBufferOverrun.. there is lots of old code that should have been rewritten long ago. Now when moving it to 'normal' it's difficult to avoid a rewrite.