1

I am using Apache Tika in .Net by converting the jar to a dll with IKVM, and running into a problem parsing images out of PDF files. Closer inspection shows this is due to some methods not being implemented in JPEGImageReader.java in the openjdk source in IKVM.

Long story short, I am looking at the TwelveMonkeys ImageIO plugin to try and work around this. However, I do not really understand how I can add this plugin when using Tika with IKVM. As I understand it, as long as the ImageIO plugins are on the class path, they should be discovered automatically.

I have tried creating a dll for Tika with the TwelveMonkeys jars as dependencies:

ikvmc.exe -target:library -version:1.15 tika-app-1.15.jar common-image-3.3.2.jar common-io-3.3.2.jar common-lang-3.3.2.jar imageio-c
ore-3.3.2.jar imageio-metadata-3.3.2.jar imageio-jpeg-3.3.2.jar imageio-tiff-3.3.2.jar

When running Tika though, this does not use the TwelveMonkeys plugins.

What's the correct way to do this?

Owen Pauling
  • 11,349
  • 20
  • 53
  • 64
  • I'm not sure it will help, even if you make it pick up the plugin, as the TwelveMonkeys `JPEGImageReader` isn't a complete re-write in Java, it still delegates to the JRE bundled `JPEGImageReader` (with its native dependencies) for the actual decoding... – Harald K Jun 15 '17 at 15:06

0 Answers0