[
https://issues.apache.org/jira/browse/TIKA-4172?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17789440#comment-17789440
]
martin k. commented on TIKA-4172:
---------------------------------
Hey [~tilman], I did read over the docs, and I should have stated this. Later
in the document (well, I read version 2.9.1), under "The default Mime Types
Detector" it also says:
{quote}Firstly, magic based detection is used on the start of the file. … Next,
if available, the filename (from TikaCoreProperties.RESOURCE_NAME_KEY) is then
used to improve the detail of the detection, such as when magic detects a text
file, and the filename hints it's really a CSV. Finally, if available, the
supplied content type (from Metadata.CONTENT_TYPE) is used to further refine
the type.{quote}
This suggests that the filename is only used to refine the detection. However,
in the case I highlighted, the content-based detection would have detected a
binary file ({{application/octet-stream}}), which then gets overruled by the
filename-based "refinement" into a {{text/*}} content type, and I think this
shouldn't be.
> Apple binary file incorrectly identified as text/x-sql due to filename
> ----------------------------------------------------------------------
>
> Key: TIKA-4172
> URL: https://issues.apache.org/jira/browse/TIKA-4172
> Project: Tika
> Issue Type: Bug
> Components: general
> Affects Versions: 2.9.1
> Reporter: martin k.
> Priority: Minor
>
> This is related to [https://github.com/eikek/docspell/issues/2376] and
> [https://github.com/eikek/docspell/issues/2403.]
> Take the following Base64 encoding of a binary Apple-generated file. No idea
> what it does. You can get the file by piping the following to e.g. {{base64
> -d > something.sql}}
> {code:java}
> ABRkMDEwMWM2Nl9teVNRTDQwLnNxbAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
> AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAbUJJTgAA
> AAAAAAAAAAAAAAAAAACCgf+/AAA=
> {code}
> If this file is name {{{}something.sql{}}}, then Tika will classify it as
> {{{}text/x-sql{}}}, which it is not. It seems like more weight is given to
> the filename (extension) than the fact that the file is binary anyway.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
