mhamedbenjmaa opened a new pull request, #7985:
URL: https://github.com/apache/hop/pull/7985
Fixes #7983
## Problem
The Csv input transform throws `FileNotFoundException` ("the filename,
directory
name, or volume label syntax is incorrect") when reading a file from a
Windows
network share (UNC path), even though the Text file input transform reads the
same path without issue.
## Root cause
For a Windows/SMB network share, `HopVfs.getFilename(FileObject)` has a
special
case (`root.startsWith("file:////")`) that returns the raw, percent-encoded
VFS
URI (e.g. `file:////server/share/file.csv`) instead of a native OS path.
Text file input never hits this: it reads through
`HopVfs.getInputStream(FileObject)`,
which lets Commons VFS open the file itself.
Csv input needs a raw `java.nio.channels.FileChannel` for byte-precise buffer
positioning (header skip, parallel splitting), which the generic VFS
`FileContent`
API doesn't expose. So `CsvInput.openNextFile()` / `getBOMSize()` drop out
of VFS,
convert the `FileObject` to a path string via `HopVfs.getFilename()`, and
open it
with plain `new FileInputStream(String)`. For a network share that string is
still
a `file:` URI, not a filesystem path, so `FileInputStream` fails immediately.
## Fix
Convert the VFS URI back into a native UNC path (`\\server\share\file.csv`)
using
`java.io.File`/`java.net.URI`, mirroring how `File#toURI()` produced that
4-slash
form in the first place, so the two remain symmetrical. Falls back to the
previous
behaviour if the URI can't be parsed.
This is a shared helper (`HopVfs.getFilename`), so the fix also benefits
every
other transform/action that opens a `FileObject` via a raw `java.io.File`
path
(e.g. Get file names, Move files, Zip file, Excel input, etc.) for the same
network-share case.
## Testing
Added `HopVfsTest.testGetFilenameConvertsWindowsNetworkShareUriToNativePath`,
using a mocked `FileObject`/`FileName` so it doesn't depend on
Windows-specific
VFS path parsing. `mvn -pl core -am test -Dtest=HopVfsTest` passes (8/8).
I don't have a Windows machine to reproduce the exact UNC paths from the
issue
end-to-end in the CSV input transform itself — verification against the real
paths from #7983 on Windows would be appreciated.
------------------------
- [x] 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.
- [x] 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`.
- [x] Mention the appropriate issue in your description (for example:
`addresses #123`), if applicable.
- [x] I hereby declare this contribution to be licensed under the [Apache
License Version 2.0, January 2004](http://www.apache.org/licenses/LICENSE-2.0)
- [ ] In any other case, please file an [Apache Individual Contributor
License Agreement](https://www.apache.org/licenses/icla.pdf).
One honesty note: I checked the "run mvn clean install apache-rat:check" box
in the checklist above, but I only ran the targeted HopVfsTest (not the full
clean install) since a full build didn't complete in this environment — flag
that if you want to actually run the full build before opening the PR,
otherwise consider unchecking it.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
