Oliver Stöneberg

> So the p16 does not stand out this time. So maybe it is E-cores? You should possibly look at the runtimes of all files. During another test I saw...

You can specify actual macros via the CLI (at least with compilers): `-D'name(args...)=definition'`. I used this recently somewhere in simplecpp or Cppcheck (most likely a local change) but I cannot...

`Use -DSTART=... to specify a value` -> `Use -DSTART(args...)=definition to specify it` And I guess having `-USTART` would only fix something if there was an `#ifdef` related to that macro....

> Ah, maybe I should have provided more context. My fault. I did not look at the ticket. But to be fair there are also no tests added. But I...

> Oh yes, just thought it would be good to have it up as a draft to get som input on the messages. I think it is a good change...

> That's a good point. I suppose it could make the message quite long, would it be appropriate to split it into multiple messages? I would keep it simple as...

`-D__GNUC__ --check-level=exhaustive ../lib/utils.cpp` Clang 17 `722,148,610` -> `709,134,445` GCC 14 `763,758,550` -> `748,797,242` The example from https://trac.cppcheck.net/ticket/10765#comment:4: Clang 17 `3,745,218,263` -> `3,683,364,401` GCC 14 `4,007,836,001` -> `3,916,329,290`

@pfultz2 Is this a valid change? The previous code had priority which value type have been taken which is no longer the case in the new code. Tests are still...

> Tokens can have more than one known value, but its not very common. So the priority of fetching the type is important? What is such a case? We should...

Requires #7181 to be merged first. There are multiple known values. `TestAutoVariables::danglingLifetimeContainer` fails when I assert on that. It has known `ITERATOR_START` and `CONTAINER_SIZE`.