[
https://issues.apache.org/jira/browse/COMPRESS-638?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17679451#comment-17679451
]
Gary D. Gregory commented on COMPRESS-638:
------------------------------------------
Percent-encoding is now used for non-ISO_8859_1 characters.
Please see git master or a snapshot build here:
https://repository.apache.org/content/repositories/snapshots/org/apache/commons/commons-compress/1.23-SNAPSHOT/
This should be considered a workaround IMO and we could change the
percent-encoding format to something else in the future if needed.
There is no roundtrip back to the non-ISO_8859_1 characters when reading a GZip
since we cannot tell what the intent of the file name bytes really is, unless
we used some special marker, which is possible in the future in suppose.
> The GzipCompressorOutputStream#writeHeader() uses ISO_8859_1 to write the
> file name and comment. If the strings contains non-ISO_8859_1 characters,
> unknown characters are displayed after decompression. Use percent encoding
> for non ISO_8859_1 characters.
> --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
>
> Key: COMPRESS-638
> URL: https://issues.apache.org/jira/browse/COMPRESS-638
> Project: Commons Compress
> Issue Type: Bug
> Reporter: Radar wen
> Priority: Major
> Fix For: 1.23
>
> Attachments: 0110.png
>
>
> The GzipCompressorOutputStream#writeHeader method uses the ISO_8859_1 to
> write the file name.
> If the file name contains non-ISO_8859_1 characters, some unknown characters
> are displayed after decompression. !0110.png!
> Can change the ISO_8859_1 to UTF-8?
> if (filename != null) {
> out.write(filename.getBytes(ISO_8859_1));
> out.write(0);
> }
>
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
