search

4teamwork / ftw.tika

This product integrates Apache Tika for full text indexing with Plone.
4 stars 1 forks source link

Add support for Tika 1.8 #27

Closed lukasgraf closed 8 years ago

lukasgraf commented 8 years ago

This PR adds support for Tika 1.8.

There have not been any necessary changes in the ftw.tika code to work with Tika 1.8, but there's a new feature we can take advantage of:

The Tika JAXRS Server in 1.8 now can return Java stack traces in the response body for failed conversions (if the -includeStack command line option was given). This allows us to do better error detection on the ftw.tika side, for example for detection conversion failures that happened because of passwort protected documents.

Therefore ftw.tika now tries to make use of any stack traces if present, but still will work just fine if the -includeStack flag is omitted (just with less detailed error logging).

If the environment variable FTW_TIKA_VERBOSE_LOGGING is set to something truthy, ftw.tika now also will log the Java stack trace (if available) upon conversion errors.

In addition, I extended the test setup so that ftw.tika is now tested against all supported versions of Tika.

@phgross @deiferni @jone

jone commented 8 years ago

:+1: cool Do we need to bumb versions in ftw-buildouts?

lukasgraf commented 8 years ago

@jone yes, we should (and probably also include -includeStack). I'm not done bumping versions in ftw.tika yet though, I'm going all the way for 1.11 ;-)

jone commented 8 years ago

:+1: :stuck_out_tongue_winking_eye: