While working on one of our projects, we came across this problem with Tomcat 5.5.9 (not sure if this exists in other versions). With compression on Text files switched on, if we are trying to download a text file, Internet Explorer (v6.0) does not recognize the file type and shows a wierd filename. Not sure how other browsers work in this situation.
Is there a solution for this without having to switch of the compression? or switching off compression is the only option? However switching off compression will be a huge overhead for the server since the application involves download of a lot of text files.
In another case, now with the compression switched off, if we are trying to download Blank Files (yes, I know this is a little wierd ), IE6 shows the same behavior as above. Whereas Mozilla (v1.0.7) does not download the file at all.