Filename with tab returned by getFileName()

Question

Our product uses JavaMail (1.4) to import mail messages from IMAP servers into our product.
One of our clients is currently having a problem that sometimes attachments are not imported because they cannot be saved due to incorrect characters in the filename. Part of the reason is that we did not sanitize the output from Part.getFileName() (a MimeBodyPart in this case)
We found out that in this case getFileName() returns a filename containing a tab (ascii 0x09), because (probably) their mailclient (Groupwise 7.0.2 HP) does not follow the intention of RFC 822 / 2822 (although it does conform to the letter of RFC 822) for the handling of long header fields. The mailclient has - as far as I can see - replaced a space in the filename by CR/LF and a tab (where I would expect CR/LF and a space).
When JavaMail decodes the header this is (of course) decoded by replacing the CR/LF + tab by a tab, and getFileName simply returns the value of the filename parameter of the Content-Disposition header (including the tab).
I would like to know:
- Is there a system property to 'fix' this behaviour
- Could you change the behaviour of getFileName() so it replaces any tabs in the filename with spaces (as far as I know there is no OS that allows tabs in filenames)

Answer 1

Hi..thanks a bunch! The quotes did help. But there was another glitch. I could not store the existing file name in $oldfilename variable as only that part of the filename before the tab got stored. So wrote the filenames in the folder into a list file and then renamed one by one. Below is the code:
for file in `ls $FilePath | sed 's/[     ]/,/g'`
do
echo $file >> $FilePath/"TabFiles"
done
ListFile=$FilePath/TabFiles
for line in `cat ${ListFile}`
do
OldFile="$(echo "$line" | sed 's/[,]/     /g')"
NewFile="$(echo "$line" | sed 's/[,]//g')"
if [ "$OldFile" != "$NewFile" ]
then
mv "$OldFile" "$NewFile"
fi
done
rm -f $FilePath/"TabFiles"
I know the idea looks rather clumsy. Is there a better way of handling this?