[
https://issues.apache.org/jira/browse/CAMEL-5495?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Aki Yoshida updated CAMEL-5495:
-------------------------------
Attachment: trunk-camel-5495.diff
patch for trunk
> file endpoints on windows without volume name in the path url not working in
> some cases
> ---------------------------------------------------------------------------------------
>
> Key: CAMEL-5495
> URL: https://issues.apache.org/jira/browse/CAMEL-5495
> Project: Camel
> Issue Type: Bug
> Components: camel-core
> Affects Versions: 2.9.2
> Environment: windows
> Reporter: Aki Yoshida
> Priority: Minor
> Attachments: trunk-camel-5495.diff
>
>
> I am using file endpoints on Windows without the volume name just like the
> normal unix based paths look.
> That means, I have something like file:///tmp/file-in and
> file:///tmp/file-out
> The funny thing that I noticed is that each endpoint itself works fine, but
> when a route is set up from one to the other, it fails because there seems to
> be an inconsistency in the way the windows path is handled in the
> camel-core's file component.
> As mentioned, each file endpoint itself is working fine. For example, the
> consumer file endpoint configured in a route
> from("file:///tmp/file-in").to("mock:test")
> works fine.
> Similarly, the produce file endpoint configured in a route
> from("direct:test").to("file:///tmp/file-out")
> works fine.
> But when a route is setup to connect these two file endpoints, the producer
> endpoint fails to create the output file.
> Concretely, there is the following code in
> GenericFileEndpoint.configureMessage that determines the file name.
> String name = file.isAbsolute() ? file.getAbsoluteFilePath() :
> file.getRelativeFilePath();
> // skip leading endpoint configured directory
> String endpointPath = getConfiguration().getDirectory() +
> getFileSeparator();
> In this particular case, the name variable is set to
> "C:\tmp\file-in\sample.xml" while the endpointPath variable is set to
> "\tmp\file-in".
> So, the subsequent code to extract the file name part, shown below, fails to
> match the path.
> if (ObjectHelper.isNotEmpty(endpointPath) &&
> name.startsWith(endpointPath)) {
> name = ObjectHelper.after(name, endpointPath);
> }
> As a result, the file name is not extracted as "sample.xml" but remain
> unchanged as "C:\tmp\file-in\sample.xml".
> Consequently, when the file producer endpoint tries to write this file in the
> file system, it tries to write a file as
> "/tmp/file-out/C:\tmp\file-in\sample.xml", resulting in an error.
> I modified FileComponent and FileEndpoint so that the endpoint path is stored
> correctly in this case to make the above extraction code find the file name
> part correctly.
> Attached is this proposed patch and a test case which runs on windows and on
> non-windows but this issue can only be demonstrated under windows.
> testRouteToFileOnly and testRouteFromFileOnly work without this patch while
> testRouteFileToFile fails without this patch on windows.
> Thanks for looking into this issue.
> Regards, Aki
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira