Forum Discussion

CLBMlx_16587's avatar
Icon for Nimbostratus rankNimbostratus
Dec 29, 2010

PDF with Kanji characters in name breaks WA

We have been trying to use the BigIP with WA to accelerate the delivery of PDF files from SharePoint over the WAN. However, we encountered problems with the double-byte files many months ago and have been trying to resolve the problem with F5. We have turned off PDF acceleration on the BigIP because of the problems. The problem occurs when PDF files have Kanji characters in the file name. If the file names are English (single byte) the problem does not occur, even if the content of the PDF contains Kanji characters.



After many months of working with F5, they are now saying that the problem is with the Adobe plug-in to Internet Explorer. Specifically they say:



The issue is a result of Adobe's (not F5's) handling of the linearization request of PDF’s with the Japanese character set over 300 KB when the Web Accelerator is enabled on the BigIP appliance. We assume the issue exists for all double-byte languages, not only Japanese. If a non-double byte character set is used, this works fine. “Linearization” is a feature which allows the Adobe web plug-in to start displaying the PDF file while it is still being downloaded in the background.



The F5 case number is: C757395.



F5 consider this an Adobe issue, not a F5 issue.



We have exhausted our contacts -- supposedly, nobody else stumbled over this Adobe bug yet, which I find hard to believe. I did post in the public Adobe forums about this as well -- not sure I'll get an answer back.



Yes, it's easy enough to tell only to use English characters, but 2/3 of our employees are overseas. We are going to constantly run into this. Haven't tried it with German Umlauts and other European special characters. Assume it would work and it's ounly double byte which doesn't.



Anybody else ran into this issue? Or has colleagues in Japan who could test this and confirm that this is indeed an universal problem?



1 Reply

  • applebee_5820's avatar
    Historic F5 Account



    I tried it and found that the error indeed occurs when a PDF file is named with Japanese charactors. I tested with code-page "Shift-JIS", but probably Unicode would produce the same result.


    I surfed Web pages open to public, and all pdf files were named with alpha-numeric charactors. However I do expect users of Intranet Web system, i.e. Sharepoint, will surely opt to use their native language for the file names.