try {
            versionRange = VersionRange.createFromVersionSpec(d.getVersion());
        } catch (InvalidVersionSpecificationException e) {
            // MNG-5368: Log a message instead of returning 'null' silently.
            this.logger.error(
                    String.format(
                            "Invalid version specification '%s' creating dependency artifact '%s'.", d.getVersion(), d),
                    e);

  public void writeUTF(String str) throws IOException {
    ((DataOutputStream) out).writeUTF(str);
  }

  // Overriding close() because FilterOutputStream's close() method pre-JDK8 has bad behavior:
  // it silently ignores any exception thrown by flush(). Instead, just close the delegate stream.
  // It should flush itself if necessary.
  @Override
  public void close() throws IOException {
    out.close();
  }

      return;
    }

    /*
     * Any of the setException() calls below can fail if the output Future is cancelled between now
     * and then. This means that we're silently swallowing an exception -- maybe even an Error. But
     * this is no worse than what FutureTask does in that situation. Additionally, because the
     * Future was cancelled, its listeners have been run, so its consumers will not hang.

          <artifactId>toolchains-maven-plugin</artifactId>
          <version>4.5.0</version>
          <executions>
            <!--
            We can apparently have only one <jdk> per execution: Others are silently ignored :(
            To properly test this, you need to remove existing toolchains:
            rm -rf ~/.m2/jdks/ ~/.m2/toolchains.xml

          <artifactId>toolchains-maven-plugin</artifactId>
          <version>4.5.0</version>
          <executions>
            <!--
            We can apparently have only one <jdk> per execution: Others are silently ignored :(
            To properly test this, you need to remove existing toolchains:
            rm -rf ~/.m2/jdks/ ~/.m2/toolchains.xml

  }

  /**
   * Invokes {@code consumer} once for each pair of <i>corresponding</i> elements in {@code streamA}
   * and {@code streamB}. If one stream is longer than the other, the extra elements are silently
   * ignored. Elements passed to the consumer are guaranteed to come from the same position in their
   * respective source streams. For example:
   *
   * <pre>{@code
   * Streams.forEachPair(

  }

  /**
   * Invokes {@code consumer} once for each pair of <i>corresponding</i> elements in {@code streamA}
   * and {@code streamB}. If one stream is longer than the other, the extra elements are silently
   * ignored. Elements passed to the consumer are guaranteed to come from the same position in their
   * respective source streams. For example:
   *
   * <pre>{@code
   * Streams.forEachPair(

 * If you use masked losses with Keras the loss values may be different in TensorFlow `2.12` compared to previous versions.

	// errors on a previous write(2) operation are reported only on the final close() that releases
	// the open file descriptor.
	//
	// Failing to check the return value when closing a file may lead to silent loss of data.
	// This can especially be observed with NFS and with disk quota.
	return w.Close()
}

// writeAllMeta - writes all metadata to a temp file and then links it to the final destination.