unresolved stack regression

[psifertex]

Between 4.1 and 4.2 we had a regression in our stack tracking for sample north wave strikes masterfully

Specifically, at offset 100258a2

The temporary work-around is to inline j__RTCCheckEsp and similar CheckEsp functions which should at least let users continue work, but we should identify whether the change that introduced the regression was an intentional fix or an accidental bug.

While this does have a work-around I'm labeling medium because I believe a regression in default analysis like this that is bad should be considered higher, but that might get changed.

[tGecko]

Hi, this also seems to affect the just uploaded sample golden pointer analyzes progressively, noticed at offset 0x00402900

[xusheng6]

For the golden pointer analyzes progressively binary, it works fine on latest dev (5.2.8366-dev Ultimate (a7397002), but it does screw up in 5.1 stable "personal 5.1.8104" as reported by one user:

This means we introduced a regression in stable 5.0, and fixed it afterwards, all without knowing 😆

This is how it looks like on latest dev, and pretty much the same as before we introduced the regression:

Note despite the presence of "unresolved stack usage", the stack variables are actually resolved fine

[xusheng6]

Also -- looking at the binary, it calls out srand as unresolved, but I checked the stack pointer and it appears to be correct, also the type library is correct