[
https://issues.apache.org/jira/browse/CAMEL-24423?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Work on CAMEL-24423 started by Andrea Cosentino.
------------------------------------------------
> camel-tika - parsed document metadata is copied to exchange headers without
> filtering
> -------------------------------------------------------------------------------------
>
> Key: CAMEL-24423
> URL: https://issues.apache.org/jira/browse/CAMEL-24423
> Project: Camel
> Issue Type: Bug
> Components: camel-tika
> Reporter: Andrea Cosentino
> Assignee: Andrea Cosentino
> Priority: Major
> Fix For: 4.23.0
>
>
> TikaProducer.convertMetadataToHeaders() copies every metadata name produced
> by the Tika parse straight onto the Camel message:
> {code:java}
> private void convertMetadataToHeaders(Metadata metadata, Exchange exchange) {
> if (metadata != null) {
> for (String metaname : metadata.names()) {
> String[] values = metadata.getValues(metaname);
> if (values.length == 1) {
> exchange.getIn().setHeader(metaname, values[0]);
> } else {
> exchange.getIn().setHeader(metaname, values);
> }
> }
> }
> }
> {code}
> The metadata names come from the parsed document, which is the message body.
> There is no HeaderFilterStrategy and no prefixing, so a document whose
> metadata declares a name inside the Camel-internal namespace sets that
> internal header on the exchange. Every other component that maps
> externally-derived names into the header map applies an inbound
> HeaderFilterStrategy for exactly this reason.
> Proposal: in convertMetadataToHeaders(), either reject names matching
> Camel*/camel*/org.apache.camel.* or prefix all parsed metadata (for example
> with "tika."), and document the choice. Prefixing is the more predictable
> option but is a breaking change for routes reading the current names, so it
> needs an upgrade-guide entry; filtering the internal namespace only is the
> smaller change.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
