[XeTeX] Bug with Unicode in hyperref PDF bookmarks

[Jonathan Kew]

On 23 Jan 2008, at 10:39 pm, James Crippen wrote:

> I'm having a Unicode problem with PDF bookmarks generated by hyperref.
> I'm running Mac OS X 10.5.1 on a MacBook Pro, with TeX Live 2007,
> XeTeX 0.997, xdvipdfmx 0.4, memoir 1.618, mempatch 4.9, fontspec 1.13,
> and hyperref 6.75r.
>
> I have some sections which have titles including the Unicode double
> quotation marks " and " in them, U+201C and U+201D. They appear fine
> in the output, but they're broken in the PDF bookmarks, instead
> appearing as â肜 (U+00E2 U+809C) for U+201C and â肝 (U+00E2 U 
> +809D) for
> U+201D. Images from Apple's Preview.app and from Adobe Acrobat are
> attached, so it's not an app specific problem. I haven't checked to
> see if the problem appears in Windows PDF viewers, but I could if
> necessary. Changing them to the proper characters in Acrobat causes no
> problems, so I presume it's something wrong somewhere between XeTeX,
> hyperref, and xdvipdfmx.
>
> Perhaps this is some sort of encoding confusion between XeLaTeX/memoir
> and hyperref? Both look suspiciously like a three byte UTF8 sequence.

Yes, they do -- but I'm a bit surprised at this behavior, as Unicode  
characters (including the double quotes -- I just tried a quick test)  
work fine for me in PDF bookmarks.

Could you provide a complete minimal example that shows this behavior  
for you, along with the .log file from the run, so we can compare  
more closely? (Try eliminating as many packages as possible -- and if  
that isolates a culprit, let us know which one.) I seem to have more  
recent versions of some packages than you listed above (I'm mainly  
running current versions from the TeX Live subversion repository),  
but am not sure what the crucial differences may be.

JK