bazelbuild/bazel · error · RuntimeError

Proguard failed ({p.returncode})

Error message

Proguard failed ({p.returncode})

What it means

Raised by the proguard wrapper after running the proguard subprocess with check=False and observing a non-zero exit code. The RuntimeError message embeds p.returncode plus any decoded stdout and stderr, so the underlying proguard diagnostics travel with the exception.

Source

Thrown at tools/build_defs/proguard/wrapper.py:101

      "-outjars",
      output_jar,
      "@" + proguard_spec,
  ]

  env = os.environ.copy()
  env.update(r.EnvVars())
  # print("Running proguard: %s" % " ".join(command))
  p = subprocess.run(command, capture_output=True, env=env, check=False)

  if p.returncode != 0:
    message = f"Proguard failed ({p.returncode})"
    stdout = p.stdout.decode()
    if stdout:
      message += f"\n  stdout:\n{stdout}"
    stderr = p.stderr.decode()
    if stderr:
      message += f"\n  stderr:\n{stderr}"
    raise RuntimeError(message)


def reset_timestamps(input_jar, output_jar, timestamp):
  """Rewrite the given jar file to reset all timestamps to a known value.

  Args:
    input_jar: The jar file to be modified.
    output_jar: The path to write the destination jar to.
    timestamp: The known timestamp to modify the output_jar with.
  """
  # print("Resetting timestamps in %s to %s, writing to %s" % (input,
  # timestamp, output))

  with zipfile.ZipFile(input_jar, mode="r") as src:
    with zipfile.ZipFile(output_jar, mode="w") as dest:
      for info in src.infolist():
        # print(f"Filename: {info.filename}")
        # print(f"  Modified: {datetime.datetime(*info.date_time)}")

View on GitHub (pinned to e6e199d060)

Solutions

1. Read the embedded stderr in the message — proguard's own diagnostics name the exact offending rule or class.
2. Add the missing jars to the deps argument so proguard can resolve referenced classes.
3. Fix or remove the offending -keep/-assumenosideeffects entry in the proguard spec.
4. If version-related, align the proguard spec dialect with the proguard binary shipped in the wrapper's runfiles.

Example fix

# before
apply_proguard(srcs=["a.jar"], deps=[], proguard_spec="spec.pro", ...)
# spec.pro: -keep class com.example.Missing { *; }

# after
apply_proguard(srcs=["a.jar"], deps=["missing.jar"], proguard_spec="spec.pro", ...)

for j in srcs + deps:
    assert os.path.exists(j), 'missing jar for proguard: %s' % j

try:
    apply_proguard(srcs, deps, spec, out)
except RuntimeError as e:
    # message embeds returncode, stdout, stderr of proguard
    log_error('proguard failure', details=str(e))
    raise

When it happens

Trigger: apply_proguard() invoked with source jars, dep jars, or a spec that proguard rejects: missing classes referenced by -keep rules, malformed .pro file syntax, duplicate class files, or incompatible proguard version.

Common situations: A -keep class references a class not in the inputs; deps list incomplete so proguard cannot resolve the classpath; spec file contains options unsupported by the shipped proguard version.

Related errors

• Invalid library proguard config parameters (these parameters
• Runfiles failed to resolve {path}
• unknown archive type "%s"
• Duplicate output file: Both {} and {} map to {}
• No <instrumentation> tag containing the targetPackage attrib

AI-assisted analysis of bazelbuild/bazel@e6e199d060 (2026-08-14). Data as JSON: /api/errors/9eb1482285033ed5. Report an issue: GitHub.