This is rather a IE-specific problem, but maybe someone has a solution. The HTTP headers are correct (Gecko for example understands and downloads/saves the file correctly) but IE does the wrong thing depending on the code:
sf = "Non-ASCII but valid filename.extension"
Response.AppendHeader("Content-Disposition", "attachment; filename=""" & sf & """")
proposes and saves under a UTF-8-encoded filename.
Response.AppendHeader("Content-Disposition", "attachment; filename=""" & Server.UrlPathEncode(sf) & """")
proposes a seemingly correct filename which, once saved on the disk, is UrlEncoded.
In both cases, if the filename is over a certain length, the filename is lost and IE proposes something random like CAC9QZGT.aspx. Changing the ContentEncoding/Charset of the response has no effect on the proposed filename.
Any ideas on how to solve this? I can't ask the users not to save the files using greek filenames...
I'd appreciate if you also cc any replies to my e-mail address.
TIA
Costas Andriotis