[
https://issues.apache.org/jira/browse/TIKA-1174?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Chris A. Mattmann updated TIKA-1174:
------------------------------------
Component/s: parser
> Invalid characters in filtered PDF output
> -----------------------------------------
>
> Key: TIKA-1174
> URL: https://issues.apache.org/jira/browse/TIKA-1174
> Project: Tika
> Issue Type: Bug
> Components: parser
> Environment: Mac OS X 10.8.5, Java 1.7u40 (but also seen on CentOS5)
> Reporter: Matt Sheppard
> Priority: Minor
> Attachments: map_sp_1c_a4.pdf
>
>
> The PDF document at
> http://www.logan.qld.gov.au/__data/assets/pdf_file/0010/9496/map_sp_1a_a4.pdf
> produces invalid characters in the output when filtered by Tika 1.4.
> {noformat}
> >
> /opt/funnelback/mbin/java/bin/java -jar tika-app-1.4.jar map_sp_1c_a4.pdf |
> hea…
> …d -n 40
> ERROR - Error: Could not parse predefined CMAP file for 'nullžf
> °-ˇžl,¡ì$1-UCS2'
> <?xml version="1.0" encoding="UTF-8"?><html
> xmlns="http://www.w3.org/1999/xhtml";>
> <head>
> [snip]
> <p>Cycle network
> </p>
> <p>
> </p>
> <p>HILEY
> </p>
> {noformat}
> Is there any proper way to avoid this, or is the best approach to strip such
> characters from Tika's output?
--
This message was sent by Atlassian JIRA
(v6.2#6252)
