Pdftk cannot work with non-ascii named files

[Joe_Bishop]

Binary package hint: pdftk

Pdftk cannot work with non-ascii named files. For example:
--------------------------------------------------------------------------------------------
master@master:~/Desktop$ ls -l
total 7112
drwxr-xr-x 3 master master 8 2007-10-28 15:04 examples
-rw-r--r-- 1 master master 5842516 2007-10-28 15:03 tutorial.tar.gz
-rw-rw-r-- 1 master master 1437044 2007-10-02 08:46 Панкратьев.pdf
master@master:~/Desktop$ pdftk Панкратьев.pdf dump_data output ahha
Error: Failed to open PDF file:
   Панкратьев.pdf
Errors encountered. No output created.
Done. Input errors, so no output created.
--------------------------------------------------------------------------------------------
It happens because input and output file names decode into Latin1 character set with the code like this one:
java::String* jv_output_filename_p= JvNewStringLatin1( output_filename.c_str() );
You can see it in the pdftk/pdftk.cc file.

[ClonedAgain (clonedagain)]

I can confirm UTF8 seems to confuse pdftk: I get "Error: Failed to open PDF file:" when I use filenames with french accents on Hardy.

Here's a blog entry from someone who claims to have a fix (source+win32 exe):
http://blog.rubypdf.com/2007/07/19/pdftk-supports-chinese-path-now/
I hav'nt had the time to try it yet, sorry.

[Adam Buchbinder (adam-buchbinder)]

The attached patch (adapted from the above link) is confirmed as fixing the issue for me. My knowledge of Debian packaging is skinny enough that I can't conjure up an actual debdiff, but I've tested adding this package via quilt and running dpkg-buildpackage; the built version does work with accented characters in filenames.

[Joe_Bishop (denis-cheremisov-gmail)]

Great! A little late, I made workaround with temporary files for this issue,
but it's better to have it late than never.

2008/9/8 Adam Buchbinder <email address hidden>

> The attached patch (adapted from the above link) is confirmed as fixing
> the issue for me. My knowledge of Debian packaging is skinny enough that
> I can't conjure up an actual debdiff, but I've tested adding this
> package via quilt and running dpkg-buildpackage; the built version does
> work with accented characters in filenames.
>
> ** Attachment added: "Patch to enable non-Latin-1 filename support."
> http://launchpadlibrarian.net/17402709/non_latin1_filename_support
>
> --
> Pdftk cannot work with non-ascii named files
> https://bugs.launchpad.net/bugs/158025
> You received this bug notification because you are a direct subscriber
> of the bug.
>

[Danilo Piazzalunga (danilopiazza)]

A simple but somewhat crude workaround would be replacing JvNewStringLatin1 with JvNewStringUTF, which assumes UTF-8. This was proposed for Debian bug #461169 (http://bugs.debian.org/461169), which affects the update_info command.

[Jose M. Albarrán (yomismo-jmalbarran)]

I confirm the same bug in spanish charset on UTF-8

[Johann Felix Soden (johfel)]

Fixed in 1.41+dfsg-1.