unsigned is a 32-bit integer (even on 64-bit platforms). File sizes are generally measured in 64-bit integers and using 32-bit integers or architecture-dependent integers is brittle and confusing.
So use u64 consistently for flatfile sizes in this refactor.
Also, use LARGE_INTEGER's QuadPart on Windows, which is recommended since C++11.
Also, unit test VarInt overflow errors.
File sizes are generally measured in 64-bit integers and using 32-bit
integers or architecture-dependent integers is brittle and confusing.
Now that C++11 added 64-bit integers, follow
https://learn.microsoft.com/en-nz/windows/win32/api/winnt/ns-winnt-large_integer-r1#remarks
> If your compiler has built-in support for 64-bit integers, use the QuadPart member to store the 64-bit integer.
Stuff like CBlockIndex::nUndoPos or CBlockFileInfo::nUndoSize is
serialized as VARINT. However, there is no test coverage at all for
VarInt overflow behavior, so add test coverage for both overflow cases
(and distinguish their error msg).
<!--e57a25ab6845829454e8d69fc972939a-->
The following sections might be updated with supplementary metadata relevant to reviewers and maintainers.
<!--006a51241073e994b41acfe9ec718e94-->
<!--021abf342d371248e50ceaed478a90ca-->
See the guideline and AI policy for information on the review process.
| Type | Reviewers |
|---|---|
| ACK | darosior, frankomosh, sedited |
If your review is incorrectly listed, please copy-paste <code><!--meta-tag:bot-skip--></code> into the comment that the bot should ignore.
<!--174a7506f384e20aa4161008e828411d-->
Reviewers, this pull request conflicts with the following ones:
util::Expected by yuvicc)If you consider this pull request important, please also help to review the conflicting pull requests. Ideally, start with the one that should be merged first.
<!--5faf32d7da4f0f540f40219e4f7537a3-->
Concept ACK
utACK fa2c8df1810f23e1c510e310eb246fd5abd3a066
ACK fa2c8df1810f23e1c510e310eb246fd5abd3a066
nPos, nSize and nUndoSize were already u64, so this makes the remaining flatfile sizes match them. I checked that the block and undo size headers on disk stay 4 bytes. Built the PR at this commit and ran serialize_tests, flatfile_tests and blockmanager_tests, which all look good.
ACK fa2c8df1810f23e1c510e310eb246fd5abd3a066