Ignore files starting with "_"

[ghost]

Dear devs,

I wanted first to thank you for this piece of software, really great!

I have one request I would like to raise with you, if possible. Could you please set the code so files starting with "" are ignored? The use case is as follows: I have a data source that is quite slow. I use Apache Flume to store that data into HDFS. Because the data velocity is small, I set up Flume to roll to a new file after 10mn. This results in creating a lot of small files which your crusher handles just perfectly. Now the issue is that Flume's temp files (i.e. files that are not closed yet) start with "" and are appended a ".tmp". When I run the crusher, if the file is closed in the meantime, well... the file is not found. I would like also to avoid errors from Flume's side and thus avoid manipulating those files.

The request is thus to either have a new option to ignore files starting with "_" or just ignore them by default.

Thanks a lot!