[llvm] Improve DWARFDebugLine::LineTable::lookupAddressRangeImpl(...) to act as exptected. (PR #226625)

Peter Rong via llvm-commits llvm-commits at lists.llvm.org
Sat Sep 26 10:26:51 PDT 2026


DataCorrupted wrote:

Overall LGTM except for nits. I wish we can change the title: "Improve" and "as exptected" is unclear. IIUC the documentation of `lookupAddressRangeImpl` didn't explicitly say whether the "expected" return value should contain Rows with identical address ranges (because when the function is written `SttmSequenceOffset` and ICF wasn't a thing yet). So I'd recommend we change that as well:

```cpp
    /// Fills the Result argument with the indices of the rows that correspond
    /// to the address range specified by \p Address and \p Size.
    ///
    /// \param Address - The starting address of the range.
    /// \param Size - The size of the address range.
    /// \param Result - The vector to fill with row indices.
    /// \param StmtSequenceOffset - if provided, only rows from the sequence
    /// starting at the matching offset will be added to the result.
    ///
    /// Returns true if any rows were found.
    bool
    lookupAddressRangeImpl(object::SectionedAddress Address, uint64_t Size,
                           std::vector<uint32_t> &Result,
                           std::optional<uint64_t> StmtSequenceOffset) const;
```

```
    /// \param Result - The vector to fill with row indices that has _unique_ ranges, duplicated ranges will be discarded (because of XYZ)
```

And the title could be "[DWARF] Update `lookupAddressRangeImpl` so the `Result` only contains unique ranges"

https://github.com/llvm/llvm-project/pull/226625


More information about the llvm-commits mailing list