MimeMessage takes too long to parse an EML (even from a local file)

Question

I am trying to parse an EML using java mail API using mimeMessage

Something like this:

InputStream stream = new FileInputStream("/my.eml");
MimeMessage message = new MimeMessage(Session.getDefaultInstance(System.getProperties()),stream);

It takes too long to read that EML if it's more than 10MBs(around 10-25 seconds)

The code is working fine so no worry about the code. I am looking something that can increase the parsing speed.

Thanks in advance.

Answer 1

Wrap your FileInputStream inside a BufferedInputStream. This way you avoid going through the storage stack of your operating system for every single byte but instead read a bunch of them in a single pass to memory.

Or for the purpose of measuring the parsing performance seperate from reading the bytes from the file try loading it in advance in memory:

    public static void main( String[] args ) throws Exception {
      File file = new File("C:\\Users\\Public\\Documents\\bigemail.eml");
      byte[] data = new byte[Math.toIntExact(file.length())];
      FileInputStream fis = new FileInputStream(file);
      int bytesRead = fis.read(data);
      System.out.println("Read " + bytesRead + " bytes");
      fis.close();

      for(int i=0; i<5; i++) {
        parseMail(new ByteArrayInputStream(data));
      }
    }

    private static void parseMail(InputStream memoryStream) throws Exception {
      var start = System.nanoTime();
      MimeMessage message = new MimeMessage(Session.getDefaultInstance(System.getProperties()),memoryStream);
      var stop = System.nanoTime();
      System.out.printf("Parsing? took %f s\n", (stop-start)/1_000_000_000.0);
      System.out.println("Subject: " + message.getSubject());
    }

On my laptop the output shows:

Read 12610414 bytes
Parsing? Took 0,261948 s
Subject: FW: .........
Parsing? Took 0,024068 s
Subject: FW: .........
Parsing? Took 0,020336 s
Subject: FW: .........
Parsing? Took 0,010707 s
Subject: FW: .........
Parsing? Took 0,010973 s
Subject: FW: .........

So the first time seems to be quite a bit slower, but for 12 MB still a far cry from what you are reporting.

This is using Oracle Java 10 on a system with 16 GB of memory and using this maven dependency:

<dependency>
  <groupId>com.sun.mail</groupId>
  <artifactId>javax.mail</artifactId>
  <version>1.6.2</version>
</dependency>

Answer 2

Try using a SharedFileInputStream .