[
https://issues.apache.org/jira/browse/TIKA-3308?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17603038#comment-17603038
]
Nick Burch commented on TIKA-3308:
----------------------------------
Our HTML mime type has both root-XML tags for well-formed documents, and a
bunch of magic for the rest. So, adding some magic as well for these documents
is in theory possible
Checking for {{<svg xmlns="http://www.w3.org/2000/svg"}} with a decent priority
should be fine, but I'm not sure we'd want to look for just {{<svg}}
Thoughts [~tallison] ?
> SVG file without xml declaration tag is detected as text/plain
> --------------------------------------------------------------
>
> Key: TIKA-3308
> URL: https://issues.apache.org/jira/browse/TIKA-3308
> Project: Tika
> Issue Type: Bug
> Components: mime
> Affects Versions: 1.25
> Reporter: Anas Hammani
> Priority: Minor
> Attachments: logo-luma.svg
>
>
> The SVG file attached to the issue is interpreted as *text/plain* by
> {code:java}
> tika.detect(filePath){code}
>
> If I add
> {code:java}
> <?xml version="1.0" standalone="no"?> {code}
> at the beginning of the file, then tika detects it as "image/svg+xml"
>
> When i read the documentation i see that xml is not necessary for a file to
> be well-formed
> [https://www.w3.org/TR/REC-xml/#sec-prolog-dtd]
>
> It will be great if tika can detect a file as a SVG without the prolog
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
