Problem/Motivation

Because Tika on Mac OS X acquires focus via the dock, indexing is disruptive to using this module on the local host. The problem is best solved on the CLI by adding the -Djava.awt.headless=true option, but this isn't directly configurable. There may be other java runtime options useful for other platforms.

Proposed resolution

The code already uses the apachesolr_attachments_tika_jar system variable, so it's just a matter of defining a form element for it and parsing the options.

Remaining tasks

finding a quick way to add whitespace between java options once they have been escaped by escapeshellarg() may introduce a security issue. Patch submitted uses a simple preg_replace, which may not be the best way to do this.

User interface changes

Adding a text box for the java CLI options in the administrative interface

EDIT: See also #1666946: Add ability to use apache tika options.

Support from Acquia helps fund testing for Drupal Acquia logo

Comments

Nick_vh’s picture

Please make a proper patch, this sounds like a reasonable fix.

wyrdwright’s picture

amontero’s picture

Adding related issue to this issue links and summary to: #1666946: Add ability to use apache tika options.