io: dlopen libbz2 by versioned SONAME, not unversioned symlink

[rgsl888prabhu]

Issue

The release/26.06 container nightly CLI test fails on the .lp.bz2 case:

----------------- CLI TEST START ---------------
Expected solution not found for .lp.bz2
Error: Process completed with exit code 1.

(e.g. https://github.com/NVIDIA/cuopt/actions/runs/26435964807/job/77827519723#step:5:1291)

Root cause

cpp/src/io/file_to_string.cpp dlopens libbz2 by its unversioned name:

dlopen("libbz2.so", RTLD_LAZY)

On Debian/Ubuntu, the unversioned libbz2.so symlink only ships with the libbz2-dev package. The runtime package (libbz2-1.0, pulled in transitively by bzip2) provides only libbz2.so.1.0 / libbz2.so.1. So dlopen returns null, the .bz2 read aborts with "Could not open .bz2 file...", and the CLI test grep miss surfaces as "Expected solution not found for .lp.bz2".

.lp.gz works in the same container because the zlib code path already uses the versioned SONAME: dlopen("libz.so.1", RTLD_LAZY).

Fix

Mirror the zlib pattern. Try names in decreasing specificity and use the first that loads:

1. libbz2.so.1.0 — bzip2 upstream's stable SONAME (objdump -p confirms; unchanged across all 1.0.x releases for ~20 years)
2. libbz2.so.1 — symlink commonly shipped alongside .so.1.0
3. libbz2.so — last resort, dev-only symlink

A future libbz2 2.x with SONAME libbz2.so.2.0 would (correctly) fail to load — it would be an ABI break anyway, same as the existing zlib behavior.

🤖 Generated with Claude Code

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: 5635aee4-2922-4f8b-b837-fa4418eb3d8d

📥 Commits

Reviewing files that changed from the base of the PR and between a995424 and a744084.

📒 Files selected for processing (1)

• cpp/src/io/file_to_string.cpp

---

📝 Walkthrough

Walkthrough

This PR enhances the dynamic loading of libbz2 in the file decompression utility. The change replaces a single attempt to load libbz2.so with a loop that tries multiple SONAME candidates (libbz2.so.1.0, libbz2.so.1, libbz2.so), improving cross-platform compatibility. The failure message is updated to report all attempted names.

Changes

bzip2 SONAME Loading Enhancement

Layer / File(s)	Summary
Multi-candidate SONAME loading for libbz2 cpp/src/io/file_to_string.cpp	Dynamic library loading now iterates through multiple libbz2 SONAME candidates instead of trying only libbz2.so, with an updated failure message listing all attempted library names.

Estimated code review effort

🎯 1 (Trivial) | ⏱️ ~3 minutes

Suggested labels

non-breaking, improvement

Suggested reviewers

• hlinsen

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title accurately and specifically describes the main change: updating bzip2 library loading to use versioned SONAMEs instead of an unversioned symlink.
Description check	✅ Passed	The description is directly related to the changeset, providing clear context about the issue, root cause, and solution being implemented in the code changes.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch fix-libbz2-dlopen-soname

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[rgsl888prabhu]

/merge