[MRESOURCES-232] Resource copy filtering should use different encoding for source and output

**[Peng Cheng](https://issues.apache.org/jira/secure/ViewProfile.jspa?name=peng)** opened **[MRESOURCES-232](https://issues.apache.org/jira/browse/MRESOURCES-232?redirect=false)** and commented

In copy-resources goal. The configuration option 'encoding' is documented to be capable of changing the encoding of source files and copy to output directory, e.g.:

```

${project.basedir}/sbin-ebcdic


${project.basedir}/sbin


IBM037

```

However, it is pointed out in the answer section of this post:

http://stackoverflow.com/questions/40095716/why-maven-resources-plugin-ignores-my-charset-encoding-setting

that this option affects both reading of source file and writing of output file, resulting in most of the file encodings not converted. This should be corrected by dividing this option into 2:

sourceEncoding & outputEncoding, which controls reading and writing separately. This is the most simple way to make encoding conversion practical.

---

**Affects:** 3.0.1

0 votes, 6 watchers

Start with the copy-resources goal and trace how its documented encoding option is applied when reading source files and writing to the output directory. Define separate sourceEncoding and outputEncoding behavior, then verify that files are converted between differing encodings without changing the existing copy-resources use case.

Written by the indexing model from the issue text.