[clang] [Analyzer][NFC] Remove redundant function call (PR #75076)

[Balazs Benics via cfe-commits]

https://github.com/steakhal commented:

I would agree with @isuckatcs, and I'd be weak against this PR.
Right now I don't see the benefit of asserting this.
Consider downstream users that might use this reporting system and have their own trackers. (We don't at Sonar, but pretend), then they would need to remove one more unjust assert.
Speaking of assert, [here](https://discourse.llvm.org/t/rfc-error-handling-in-release-builds-aka-can-i-use-lldbassert-or-not/74738/10) is my statement about them for the whole llvm prohect.
We are on the cold path anyways here. It should not matter.

https://github.com/llvm/llvm-project/pull/75076