[
https://issues.apache.org/jira/browse/COMPRESS-459?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16539618#comment-16539618
]
ASF GitHub Bot commented on COMPRESS-459:
-----------------------------------------
Github user bodewig commented on the issue:
https://github.com/apache/commons-compress/pull/67
Thanks @ctron!
Don't we need to adjust `getHeaderPadCount` in `readOldBinaryEntry` as well?
At first I stumbled over the `-1` in the call to `getHeaderPadCount` but I
think it is because the method already accounts for the trailing NUL, I guess a
comment could help :-)
One more nit, you removed a since tag from one of `CpioArchiveEntry`'s
constructors by accident.
Also, but this probably is a separate issue: `CpioArchiveOutputStream`
seems to have the same problem (even worse, as the length we write into the
header may be wrong). It seems we should deprecate the no-arg version of
`getHeaderPadCount` as it only works for single bye encodings.
> CPIO fails decoding multibyte name entries
> ------------------------------------------
>
> Key: COMPRESS-459
> URL: https://issues.apache.org/jira/browse/COMPRESS-459
> Project: Commons Compress
> Issue Type: Bug
> Components: Compressors
> Affects Versions: 1.9, 1.17
> Reporter: Jens Reimann
> Priority: Major
>
> Having a CPIO archive in (e.g. UTF-8) mode and having a name entry with a
> name containing multi-byte characters the decoder fails.
> The problem IMHO is the "getHeaderPadCount" method, which assumes a single
> byte per character:
>
> {code:java}
> public int getHeaderPadCount(){
> if (this.alignmentBoundary == 0) { return 0; }
> int size = this.headerSize + 1; // Name has terminating null
> if (name != null) {
> size += name.length();
> }
> final int remain = size % this.alignmentBoundary;
> if (remain > 0){
> return this.alignmentBoundary - remain;
> }
> return 0;
> }
> {code}
> However this may (or may not) be true for UTF-8.
>
> Also it wouldn't be enough to call "String#getBytes(…)" as this might already
> transform the underlying bytes.
> The proper solution would be to provide the name size, as read from the CPIO
> stream, and pass it to the entry.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
