[cfe-dev] [analyzer] Could you please give a descriptions to some checker options?

Artem Dergachev via cfe-dev cfe-dev at lists.llvm.org
Thu Apr 18 20:04:35 PDT 2019

On 4/18/19 10:54 AM, Kristóf Umann wrote:
> Hi!
> Just a quick question, I'm planning to land 
> https://reviews.llvm.org/D57855: "[analyzer][NFC] Reimplement checker 
> options", as it seems like there is a general consensus in the 
> community that it's looking great. However, some checker option 
> descriptions are missing, and I couldn't manage to figure them out on 
> my own, could you help me with that please?
>   * nullability:NoDiagnoseCallsToSystemHeaders

That's what the comment says, i guess i can't add much to it:

     88   // If true, the checker will not diagnose nullabilility issues 
for calls
     89   // to system headers. This option is motivated by the 
observation that large
     90   // projects may have many nullability warnings. These projects may
     91   // find warnings about nullability annotations that they have 
     92   // added themselves higher priority to fix than warnings on 
calls to system
     93   // libraries.

I.e., something like
Suppresses warnings for violating nullability annotations of system 
header functions. This is useful if you are concerned with your custom 
nullability annotations more than with following nullability 
specifications of system header functions.
>   * osx.cocoa.RetainCount:CheckOSObject

Find violations of retain-release rules applied to XNU OSObject 
instances. By default, the checker only checks retain-release rules for 
Objective-C NSObject instances and CoreFoundation objects.
>   * osx.cocoa.RetainCount:leak-diagnostics-reference-allocation

This one was removed in r348637 :)

>   * osx.cocoa.RetainCount:TrackNSCFStartParam

Check not only that the code follows retain-release rules with respect 
to objects it allocates or borrows from elsewhere, but also that it 
fulfills its own retain count specification with respect to objects that 
it receives as arguments.

> Cheers!
> Kristóf

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-dev/attachments/20190418/b43d6030/attachment.html>

More information about the cfe-dev mailing list