Fail the macOS build if any bundled library links outside the bundle#10135
Merged
Conversation
The macOS appbundle is meant to be self-contained: every binary should reference only OS libraries (/usr/lib, /System) or bundle-relative paths (@loader_path/@rpath/@executable_path). An absolute install-name pointing at a build-host location such as Homebrew's /usr/local/opt/openssl@3 or the Jenkins workspace will resolve on the build machine but is absent (or, for a differently-signed dylib, rejected by hardened-runtime library validation) on an end user's Mac, so the app dies on startup. This is exactly how 9.16 broke on Intel (issue pgadmin-org#10123): with no Intel cryptography wheel published upstream, pip compiled it from source and its openssl-sys build linked the build host's Homebrew OpenSSL into _rust.abi3.so, which _fixup_imports deliberately skips, so the dangling reference shipped unnoticed. Add a _verify_bundle_linkage step that runs after the bundle is fully assembled and relocated, and before code-signing, walking every .so/.dylib and failing the build if anything links to a build-host prefix. This turns the whole class of problem from a silent runtime failure on user machines into a hard build failure on the builder.
|
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Path: .coderabbit.yaml Review profile: CHILL Plan: Pro Plus Run ID: 📒 Files selected for processing (1)
🚧 Files skipped from review as they are similar to previous changes (1)
WalkthroughAdds ChangesBundle Linkage Verification
Estimated code review effort: 2 (Simple) | ~10 minutes Possibly related PRs
🚥 Pre-merge checks | ✅ 4 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (4 passed)
✨ Finishing Touches🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@pkg/mac/build-functions.sh`:
- Around line 498-508: The Mach-O dependency scan in the bundle check is too
narrow because the loop only inspects files matched by the .so and .dylib
suffixes, so shipped executables and framework binaries can be missed. Update
the file discovery in the build-functions.sh scan to target all Mach-O files
under BUNDLE_DIR, then keep the existing otool -L, grep, and error reporting
logic in the same loop so any host-linked binary is caught regardless of
extension.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
Run ID: e48e57f7-c8e9-440b-bd04-fe6b4372a21a
📒 Files selected for processing (2)
pkg/mac/build-functions.shpkg/mac/build.sh
asheshv
reviewed
Jul 24, 2026
asheshv
added a commit
that referenced
this pull request
Jul 24, 2026
…#10189) _verify_bundle_linkage's host-library scan only matched files by .so/.dylib name suffix, so a shipped executable (Contents/MacOS/*) or a Python.framework payload with a host-linked dependency but no matching suffix could slip through unchecked — the same class of bug #10135 was added to catch. Detect Mach-O executables/libraries by content via `file` instead (same pattern the codesign step in this file already uses), regardless of extension or executable permission bits, so readable non-executable dylibs are covered too.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Belt-and-braces guard for the macOS appbundle build.
Background
9.16 failed to start on Intel Macs (#10123). The
cryptographypackage no longer publishes a prebuilt wheel for Intel macOS, so pip compiled it from source on the builder, and itsopenssl-sysbuild linked the build host's Homebrew OpenSSL (/usr/local/opt/openssl@3/...) into_rust.abi3.sorather than the OpenSSL we ship inside the bundle._fixup_importsdeliberately skips_rust.abi3.so, so that dangling reference shipped unnoticed and the app died on startup on every machine that lacked (or could not validate) that external dylib.The actual fix lives in the build configuration (force
cryptographyto link the bundled OpenSSL). This PR is the separate safety net so the same class of problem cannot silently reach a release again.Change
Adds a
_verify_bundle_linkagestep, run after the bundle is fully assembled and relocated and before code-signing, that walks every.so/.dylibin the bundle and fails the build if any install-name points at a build-host prefix (/usr/local,/opt/homebrew,/opt/local, or$SLAVE_HOMEon the CI builders). Legitimate OS libraries (/usr/lib,/System) and bundle-relative references (@loader_path/@rpath/@executable_path) pass untouched.A correct, self-contained bundle passes silently; a left-behind external reference turns into a hard build failure on the builder instead of a runtime failure on a user's Mac.
Testing
macOS build scripts only; no runtime code paths affected.
bash -nclean on both files, and the detection pipeline was checked against the realotool -Loutput from the #10123 reports (flags the twoopenssl@3lines, ignores the@rpathself-id and the/usr/libsystem libs).Summary by CodeRabbit