[
https://issues.apache.org/jira/browse/TIKA-4356?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17900057#comment-17900057
]
Subbu commented on TIKA-4356:
-----------------------------
I am able to reproduce this issue. Can I pick this to analyse and possibly
provide a PR?
> tika core content detection possible regression for subtypes
> -------------------------------------------------------------
>
> Key: TIKA-4356
> URL: https://issues.apache.org/jira/browse/TIKA-4356
> Project: Tika
> Issue Type: Bug
> Components: core
> Affects Versions: 3.0.0
> Reporter: Francesco
> Priority: Major
>
> Hi,
> after updating Tika core from 2.9.2 to 3.0.0.
> Seems tika is not able anymore to recognize subtypes with only glob pattern,
> like application/json
> triying to detect a json like \{"test": "test"}, specifing file name
> "file.json" the result from tika is text/plain
> while the older version recognize correctly the application/json mime.
> the following code works with the 2.9.2 version and return text/plain with
> the 3.0.0
> {code:java}
> Tika tika = new Tika();
> String jsonString ="{\"test\": \"test\"}";
> String result = tika.detect(jsonString.getBytes(), "test.json");{code}
> Has something changed and now i'm doing it wrong or it is a regression?
>
> thanks.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
