Oliver Stöneberg
Oliver Stöneberg
Discovered by assuming that https://trac.cppcheck.net/ticket/12125 and https://trac.cppcheck.net/ticket/10949 were fixed which indeed they were not.
CC @danmar @orbitcowboy
IMO `exhaustive` need to be the default so somebody running this the first time does not get any warnings about this. If they have issues with the run-time then they...
Disabling it by default also completely killed the `valueFlowBailoutIncompleteVar` data collection in daca: http://cppcheck1.osuosl.org:8000/value_flow_bailout_incomplete_var.html.
I also think we should rename the check level `normal` to `reduced` so it better reflects what it actually does. And the current `exhaustive` is also a bit misleading as...
Sorry for the late reply but I wanted to finish up some of my other stuff before I dive into this but that didn't work out as well as I...
> Disabling it by default also completely killed the `valueFlowBailoutIncompleteVar` data collection in daca: Restored in #6153.
#6153 also revealed some things about the ValueFlow I was not (actively) aware of. We perform it for everything no matter if it is used or not. It seems if...
We are treating certain errors as critical because they cause the analysis of subsequent code to stop. You could argue that any bailout is similar to that. Be it either...
The reduced default check level might also cause misleading results in the tests. We should default at exhaustive in the object but only set the different default when parsing the...