That's a pretty wonky experiment, even if a fairly easy one. I'm trying to add support for fix-it hints to the bug reporter. In the current shape the patch does the following:
- Allow attaching fixit hints to Static Analyzer BugReports.
- Add support for fixits in text output (including the default text output that goes without notes, as the fixit "belongs" to the warning).
- Add support for fixits in the plist output mode (not sure if i'm fully supporting all kinds of fixits).
- Implement a fixit for the path-insensitive DeadStores checker (only one of the cases, and i'm still not sure it's a good fixit, but it was an obvious first thing to experiment with).
- Implement a fixit for the path-sensitive VirtualCall checker when the virtual method is not pure virtual (in this case the "fix" is to suppress the warning by qualifying the call).
$ clang -cc1 -analyze -analyzer-checker=core,cplusplus,optin.cplusplus test.cpp test.cpp:4:5: warning: Call to virtual method 'S::foo' during construction foo(); ^~~~~ S:: 1 warning generated.
- We don't have a good way to test removal fixits (such as the one in the dead stores checker). Testing plist files is not a good way to test them (though we definitely need a few such tests). We can't test them via text output because clang itself generally doesn't display removal fixits in text output (it's kinda obvious if you look at how it prints them). In clang-tidy they use a more sophisticated facility for these tests, testing the fixed file instead, but it's deep within their custom testing scripts. We might need something similar.
- HTML output still doesn't support fixits. Not sure if they are useful in there because fixits are generally not very useful when there's no button to apply them. But there should be no harm in displaying them anyway, so i hope i'll have time to take a look.
- Need more tests with macros and such stuff.