Sean Palmer created RAT-582:
-------------------------------
Summary: Scripts, batch files, and DTDs are never checked
Key: RAT-582
URL: https://issues.apache.org/jira/browse/RAT-582
Project: Apache RAT
Issue Type: Bug
Components: mime-meta-data
Affects Versions: 0.18
Reporter: Sean Palmer
Files whose Tika media type is in the {{application/...}} tree are classified
as BINARY unless listed in {{TikaProcessor.DOCUMENT_TYPE_MAP}}, so their
license headers are never checked. This affects several common source formats:
* {{.sh}}, {{.bash}} ({{application/x-sh}})
* any file starting with a {{#\!/}} line whose interpreter Tika doesn't
recognise by path, e.g. {{#\!/usr/bin/env python3}}, {{#\!/usr/bin/env perl}},
{{#\!/usr/bin/env node}}. Tika's {{application/x-sh}} magic on {{#\!/}} wins
over the file name, so a {{.py}} file becomes {{application/x-sh}}
* {{.bat}}, {{.cmd}} ({{application/x-bat}})
* {{.dtd}} ({{application/xml-dtd}})
By contrast, {{#\!/usr/bin/python}} is detected as {{text/x-python}} and is
checked.
Found while investigating
https://github.com/apache/tooling-trusted-releases/issues/1725. Apache Trusted
Releases runs RAT on uploaded source archives.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
