Skip to content

fix(rest): enforce path encoding - #16580

Open
wcornell wants to merge 2 commits into
googleapis:mainfrom
wcornell:fix-rest-path-encoding
Open

wcornell wants to merge 2 commits into
googleapis:mainfrom
wcornell:fix-rest-path-encoding

Conversation

@wcornell

Copy link
Copy Markdown

Generated REST stubs previously constructed request URI paths with absl::StrCat, without percent-encoding path parameters or validating against . / .. traversal segments, unexpected / characters in single-segment fields, or malformed UTF-8.

  • Update internal::UrlEncode to strictly follow the RFC 3986 Section 2.3 unreserved character set
  • Add rest_internal::EncodePath / rest_internal::PathPiece to validate and encode single-segment, multi-segment (=**), and custom-verb (:verb) path components (returning StatusCode::kInvalidArgument on invalid inputs)
  • Update the generator and Discovery proto converter to emit EncodePath across all generated REST stubs.

The first commit contains all library, generator, golden, and test changes; the second commit contains only the regenerated Discovery .proto and *_rest_stub.cc files.

@wcornell wcornell changed the title Fix REST path encoding fix(rest): Enforce path encoding Oct 20, 2026
@wcornell wcornell changed the title fix(rest): Enforce path encoding fix(rest): enforce path encoding Oct 20, 2026

@gemini-code-assist gemini-code-assist Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Code Review

This pull request refactors REST path construction across generated stubs by replacing absl::StrCat with rest_internal::EncodePath and rest_internal::PathPiece. This change ensures that path segments are safely encoded and validated, rejecting invalid segments containing . or .. with an argument error. The generator, along with its unit and integration tests, has been updated to support and verify this new path encoding behavior. I have no further feedback to provide as there are no review comments to assess.

@codecov

codecov Bot commented Oct 20, 2026 •

Copy link
Copy Markdown

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 92.42%. Comparing base (513468f) to head (4c7a4d9).

Additional details and impacted files
@@            Coverage Diff             @@
##             main   #16580      +/-   ##
==========================================
+ Coverage   92.39%   92.42%   +0.03%     
==========================================
  Files        2266     2268       +2     
  Lines      220060   220564     +504     
==========================================
+ Hits       203317   203852     +535     
+ Misses      16743    16712      -31     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:
  • ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

@wcornell
wcornell force-pushed the fix-rest-path-encoding branch from 9b076c5 to 4c7a4d9 Compare October 23, 2026 20:58
@wcornell
wcornell marked this pull request as ready for review October 26, 2026 16:16
@wcornell
wcornell requested a review from a team as a code owner October 26, 2026 16:16
GOOGLE_CLOUD_CPP_INLINE_NAMESPACE_BEGIN
namespace {

bool IsInvalidSingleSegment(std::string_view value) {

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can we refactor this from IsInvalidSingleSegment to IsValidSingleSegment? This avoids having to reason about double negatives.

} // namespace

StatusOr<std::string> EncodePathImpl(std::initializer_list<PathPiece> pieces) {
if (pieces.size() == 0) return std::string("/");

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It pains me that std::initializer_list::empty was not added until C++26.

}

TEST(UrlEncode, HighBitBytes) {
absl::string_view constexpr kUnencodedString = "\x80\xC3\xA9\xFF";

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Let's make sure we use std::string_view instead of absl::string_view here and anywhere else in this PR.

We should probably update the AGENTS.md file to call this (and other absl types that we now have access to in C++17) out.

This branch was previously deployed

1 inactive deployment
false — 4c7a4d93 Deployed Oct 23, 2026 by wcornell via Save PR ref #12600
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants