[jira] [Commented] (TIKA-1141) javascript files that contain "

[Daniel Goltz (JIRA)]

    [ 
https://issues.apache.org/jira/browse/TIKA-1141?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16299893#comment-16299893
 ] 

Daniel Goltz commented on TIKA-1141:
------------------------------------

[~gagravarr] this is our soltution to this issue:
{code:java}
private String detectContentType(FilePath file) throws InterruptedException, 
IOException {
        Tika tika = new Tika();
        try (InputStream stream = file.read()) {
            if (file.getName().endsWith(".js")) {
                // Tika has a shortcoming not able to properly identify 
JavaScript files, determine type by extension
                // rather than Tika for those.
                return "application/javascript";
            } else {
                return tika.detect(stream, file.getName());
            }
        }
    }
{code}
As you can see, the file name exists and is definitly pased to Tika otherwise 
our solution wouldn't have worked.
We are using a jquery version that was minified in-house but this should have 
worked in any case if the file name is passed.

> javascript files that contain "<html" are detected as text/html
> ---------------------------------------------------------------
>
>                 Key: TIKA-1141
>                 URL: https://issues.apache.org/jira/browse/TIKA-1141
>             Project: Tika
>          Issue Type: Bug
>          Components: mime
>    Affects Versions: 1.2
>            Reporter: David Hara
>            Priority: Minor
>
> The Mimetypes detector will return text/html as the mimetype for any 
> javascript file that contains the string "<html" in it. I believe this is due 
> to the rule <match value="&lt;html" type="string" offset="0:8192"/> in the 
> tika-mimetypes.xml file.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)