Spec cleanup: Teredo, record sizes, and deprecated types#220
Conversation
The previous wording ("stored as-is ... first 32-bits of the address
space") was ambiguous and led implementers to incorrectly place IPv4
bytes in the most significant bits of the 128-bit address. Rewrite to
explicitly state that IPv4 bytes go in the least significant 32 bits,
and describe the tree traversal implications.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Summary of ChangesHello @oschwald, I'm Gemini Code Assist1! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed! This pull request enhances the clarity and accuracy of the MaxMind DB specification by providing more precise details on how various IP address types are handled within the database structure. It refines the definition of record sizes, formally deprecates unused data types, and includes several editorial improvements. The changes aim to improve understanding for implementers and maintainers of the specification. Highlights
Changelog
Activity
Using Gemini Code AssistThe full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips. Invoking Gemini You can request assistance from Gemini at any point by creating a comment using either
Customization To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a Limitations & Feedback Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here. You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension. Footnotes
|
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request introduces a set of clarifications and cleanups to the MaxMind DB file format specification. The changes address several areas: the handling of Teredo subnets is explained in greater detail, the constraints on record sizes are defined more precisely, and the unused data types 12 (data cache container) and 13 (end marker) are formally deprecated while preserving historical context. The pull request also includes various editorial improvements, such as clarifying how IPv4 addresses are represented in IPv6 trees and fixing minor grammatical errors and inconsistencies. The changes improve the overall clarity and accuracy of the specification. I have reviewed the modifications and have not identified any issues.
oschwald
force-pushed
the
greg/eng-4268
branch
from
191b574 to
1affb1d
Compare
- Add missing "is" in bytes data type description - Add missing "that" in 28-bit node layout note - Change "comprised of" to "composed of" in overview Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The data type is called "map" everywhere else in the spec. This was leftover terminology. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
These labels were trying to distinguish inline types (1-7) from extended types (8+), but both are binary encodings. The distinction is already explained in the surrounding text. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Simplify "the next byte after the type specifying bytes as an unsigned integer" to "the next byte (as an unsigned integer)" in all three size extension descriptions. The phrase "after the type specifying bytes" is unnecessary since the description already establishes sequential reading. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Explain why a tree alias cannot resolve Teredo client IPv4 addresses: the client IP is in bits 96-127, XOR'd with 0xFFFFFFFF, and separated from the Teredo prefix by 64 bits of unrelated data. A tree alias can only map bits immediately following the aliased prefix. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Replace the misleading "24 to 128 bits" range with precise constraints: record size must be a multiple of 4, at least 24 bits. Note that all existing databases use 24, 28, or 32 bits while the format supports larger sizes. Drop the redundant "(24, 28, or 32)" parenthetical from the Search Lookup Algorithm section. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Neither type has ever been produced by any writer or appeared in any known database. Type 12 (data cache container) is superseded by pointer- based deduplication. Type 13 (end marker) is superseded by metadata- based section boundary calculation. Mark both as deprecated and explain their original intent. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Wraps markdown prose at 80 characters using prettier's --prose-wrap always option. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Wrap prose at 80 characters using prettier --prose-wrap always, matching the new prettier-markdown command in .precious.toml. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
oschwald
force-pushed
the
greg/eng-4268
branch
from
1affb1d to
cc5279e
Compare
2 participants
Summary
0xFFFFFFFF, and separated from the prefix by 64 bits of unrelated data. Add RFC 4380 reference.Also includes earlier editorial fixes: IPv4-in-IPv6 tree storage clarification, typo/grammar fixes, "hash" → "map" terminology, ASCII/binary label removal, and payload size phrasing.
Test plan
precious lint --allpasses after each commit🤖 Generated with Claude Code