Issue 887 in lilypond: PDF point-and-click URI links with special chars wrongly encoded depending on the OS

[codesite-noreply]

Status: Accepted
Owner: v.villenave
Labels: Type-Defect Priority-Medium OpSys-Linux OpSys-OSX

New issue 887 by v.villenave: PDF point-and-click URI links with special  
chars wrongly encoded depending on the OS
http://code.google.com/p/lilypond/issues/detail?id=887

New report from Dénes:
http://lists.gnu.org/archive/html/bug-lilypond/2009-09/msg00090.html

Point-and-click links in Lily-generated PDF files do not work if they link
to a file with a non-ASCII filename or path.  The PDF spec says that the
URIs should be encoded as 7-bit ASCII, instead they are encoded in the
system's default encoding (Cp1250 on Windows, UTF-8 on Unixes) and then
escaped with backslash-codes.

Here's a not-for-the-faint-hearted example:

Links to a file named árvíztűrő tükörfúrógép.ly look like this:

 - On Windows:
C:\Documents and Settings\Harmath
Denes\workspace-classic\.runtime\árvíztűrő tükörfúrógép\árvíztűrő
tükörfúrógép.ly ->
textedit://C:/Documents%20and%20Settings/Harmath%20Denes/workspace-classic/.runtime/\341rv\355zt\373r\365%20t\374k\374rf\372r\363g\351p/\341rv\355zt\373r\365%20t\374k\366rf\372r\363g\351p.ly

This is correct, the system finds the file.

 - But on OS X...:
/Users/thsoft/Development/.runtime/árvíztűrő tükörfúrógép/árvíztűrő
tükörfúrógép.ly ->
textedit:///Users/thsoft/Development/.runtime/a\314\201rvi\314\201ztu\314\213ro\314\213%20tu\314\210ko\314\210rfu\314\201ro\314\201ge\314\201p/\303\241rv\303\255zt\305\261r\305\221%20t\303\274k\303\266rf\303\272r\303\263g\303\251p.ly

 - and on GNU/Linux:
textedit:///home/valentin/\303\241rv\303\255zt\305\261r\305\221%20t\303\274k\303\266rf\303\272r\303\263g\303\251p/\303\241rv\303\255zt\305\261r\305\221%20t\303\274k\303\266rf\303\272r\303\263g\303\251p.ly

Notice that in this case, accented characters in the path (e.g. á ->
a\314\201) are encoded differently (!) than accented characters in the
filename (á -> \303\241)...
Seems like that the latter is correct, but only with Latin-2 characters;
further experiments showed that:
a\á.ly is correct
a\ő.ly is incorrect (ő is not in the Latin-2 charset)
á\*.ly are always incorrect

--
You received this message because you are listed in the owner
or CC fields of this issue, or because you starred this issue.
You may adjust your issue notification preferences at:
http://code.google.com/hosting/settings