Loading external vs internal DWARF results differ

[negasora]

Repro steps:

1. Open test_stripped
2. Import debug info from test_debuginfo
3. Open test_unstripped
4. Navigate to main in both and notice that the stack layout is different

dwarf_bug.zip

There are a handful of log errors that make me think that we're adding the bv start address twice somewhere

[Analysis] Attempting to add function not backed by file: 0x8021d0
[Analysis] Attempting to add function not backed by file: 0x802230
[Analysis] Attempting to add function not backed by file: 0x802290
[Analysis] Attempting to add function not backed by file: 0x802220
[Analysis] Attempting to add function not backed by file: 0x802350
[Analysis] Attempting to add function not backed by file: 0x802200
[Analysis] Attempting to add function not backed by file: 0x802340
[Analysis] Attempting to add function not backed by file: 0x8022c0
[Analysis] Attempting to add function not backed by file: 0x802310
[Analysis] Attempting to add function not backed by file: 0x8022b0

[negasora]

Part of this was because there is a case where we're adding the binary base to function addresses twice, and the other part of this is because the CIE offsets are stored in the non-debug binary instead of the debuginfo binary, and we're refusing to create local variables without known CIE offsets

[negasora]

This is fixed on the test_fix_dwarf_bugs_2025_4_18 branch, probably to be merged after 5.0 stable releases

[negasora]

Fixed in 5.1.7295-dev