[
https://issues.apache.org/jira/browse/TIKA-4882?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18112978#comment-18112978
]
ASF GitHub Bot commented on TIKA-4882:
--------------------------------------
tballison commented on PR #3142:
URL: https://github.com/apache/tika/pull/3142#issuecomment-5590936152
Thank you for opening this!
In looking at the codebase, I realize that we let duplicative call sites
arise, which we shouldn't have.
EmbeddedDocumentUtil.getExtensionForMediaType(String) already basically does
what FileNameUtils does, with some difference input and default missing value.
We should put this functionality in one, along the lines of:
```
public static String calculateExtension(Metadata metadata, String
defaultValue) {
String ext = EmbeddedDocumentUtil.getExtensionForMediaType(
metadata.get(HttpHeaders.CONTENT_TYPE));
return StringUtils.isBlank(ext) ? defaultValue : ext;
}
```
Perhaps on a follow on PR, put `public String getExtension(String name) {}
in MimeTypes and then call from there?
Consider deprecating the two current call sites that are nearly duplicative?
> FilenameUtils.calculateExtension ignores defaultValue on unknown path
> ---------------------------------------------------------------------
>
> Key: TIKA-4882
> URL: https://issues.apache.org/jira/browse/TIKA-4882
> Project: Tika
> Issue Type: Bug
> Components: core
> Reporter: Tim Grein
> Priority: Minor
>
> FilenameUtils.calculateExtension ignores an unknown extension in the sense
> that the behavior on an unknown extension is not specified in the javadoc and
> it's not explicitly handled. It should probably be either handled or
> explicitly documented.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
