C++: Fix typeid IR translation

[jketema]

No description provided.

[Copilot]

Pull Request Overview

This PR fixes the IR translation for C++ typeid expressions by implementing proper IR instructions for both typeid with expressions and typeid with types. The changes address consistency issues in the IR generation where typeid operations were previously not properly handled.

Key changes:

• Adds new IR instruction types TypeidExpr and TypeidType with corresponding opcodes
• Implements TranslatedTypeidExpr class to handle the translation of typeid operations
• Updates test expectations to reflect the corrected IR generation

Reviewed Changes

Copilot reviewed 10 out of 10 changed files in this pull request and generated no comments.

File	Description
cpp/ql/lib/semmle/code/cpp/ir/implementation/Opcode.qll	Adds TTypeidExpr and TTypeidType opcodes with their corresponding classes
cpp/ql/lib/semmle/code/cpp/ir/implementation/raw/Instruction.qll	Defines TypeidExprInstruction and TypeidTypeInstruction classes
cpp/ql/lib/semmle/code/cpp/ir/implementation/raw/internal/TranslatedExpr.qll	Implements TranslatedTypeidExpr class for handling typeid expression translation
Multiple .expected files	Updates test expectations to reflect the fixed IR translation

[MathiasVP]

One comment, but otherwise this LGTM if DCA comes back happy!

[MathiasVP]

Final rounds of comments and then I think we're there 🥳

[MathiasVP]

LGTM (once DCA comes back happy)!

[jketema]

It looks likes this gives a slowdown on zeek/spicy. It looks like that project uses typeid in some of it's general error handling routines, so I suspect we get a lot more flow through those now. It's difficult to exactly tell from the DCA data though.

The other projects in DCA look fine.

[MathiasVP]

Makes sense. I think that's fine. I'm quite happy with the large reduction in instructionWithoutSuccessor 🥳

Did you see that there's a large increase in missingOperandType consistency errors for drogon?

[jketema]

Did you see that there's a large increase in missingOperandType consistency errors for drogon?

Yes, drogon also has some uses of typeid (including some in templates), so I assume it has to do with that. Not sure if I should investigate here.

[MathiasVP]

Yeah, I don't think this needs to be looked at in this PR since I highly doubt that this PR does anything incorrectly.

[jketema]

Ok, merging.