[bamaer]

…ath in SFTP

fixes #9033

Please add a meaningful description for your change here

---

Thank you for your contribution! Follow this checklist to help us incorporate your contribution quickly and easily:

• Run mvn clean install apache-rat:check to make sure basic checks pass. A more thorough check will be performed on your pull request automatically.
• If you have a group of commits related to the same change, please squash your commits into one and force push your branch using git rebase -i.
• Mention the appropriate issue in your description (for example: addresses #123), if applicable.

To make clear that you license your contribution under the Apache License Version 2.0, January 2004
you have to acknowledge this by using the following check-box.

• I hereby declare this contribution to be licensed under the Apache License Version 2.0, January 2004
• In any other case, please file an Apache Individual Contributor License Agreement.

[hansva]

1. Must fix: SftpClient still sends //X as is. chdir(), folderExists(), createFolder() (and getFileType(), delete(), renameFile() with absolute paths) don't add the trailing slash, so jsch still turns a single-level data set path into /X for ActionSftp and ActionSftpPut. Move the rule into one helper and use it in SftpFileObject.remotePath and in those methods:

public static String keepDataSetPrefix(String path) {
  return path != null && path.length() > 2 && path.startsWith("//")
          && path.charAt(2) != '/' && path.indexOf('/', 2) < 0
      ? path + "/"
      : path;
}

Add SftpClient cases to SftpDoubleSlashPathTest. If you'd rather leave this out of the PR, change "fixes #9033" to "addresses #9033" and open a follow-up issue.

2. Must fix: PR description. It's still the template text. Add the cause, the fix, and what stays open from #9033: the hint for invalid data set names, and the jsch upstream report.

3. Should do: check a sequential data set. Every single-level path now ends in /, files included. Read and write one single-level sequential data set on z/OS. If //HLQ.SEQ/ is refused, say in the comment on remotePath that only folders work.

4. Nit: add the z/OS sentence of SftpPut.Error.TargetIsNotAFile only for // paths (targetFile.getName() instanceof IRootPrefixFileName p && !p.getRootPrefix().isEmpty()).

[mattcasters]

[suggestion] keepDataSetPrefix appends / to every single-level // path, and SftpFileObject.remotePath plus SftpClient stat/delete/rename/cd all use it. jsch then sends //HLQ.SEQ/ for a sequential data set opened as a file, not only for a prefix used as a folder. The javadoc says that file form has not been verified on z/OS; a trailing slash is the directory/prefix spelling, so a read, write, or rm can miss //HLQ.SEQ or address a different name. Previously glob turned it into /HLQ.SEQ, so this is not a regression of a working path, but it is a different path.

Suggestion: Read and write one single-level sequential data set on z/OS. If //HLQ.SEQ/ is accepted, say so here. If it is refused, say in this comment that only a folder/prefix is covered, and do not treat file get/put/rm of //HLQ.SEQ as fixed.

[mattcasters]

[nit] The z/OS sentence is appended for every non-file target, including an ordinary directory on a Unix SFTP server.

Suggestion: Keep the first sentence for every target. Add the z/OS clause only when the name is a // path (targetFile.getName() is an IRootPrefixFileName with a non-empty root prefix).

[mattcasters]

[nit] Root cause: is a new string shown on the error row, and it is hardcoded in English while the other SFTP Put errors go through the message bundle.

Suggestion: Move it to messages_en_US.properties (for example SftpPut.Error.RootCause=Root cause: {0}) and pass ExceptionUtils.getRootCauseMessage(e) as the argument.