Page MenuHomePhabricator

PdfHandler doesn't display all unicode characters properly
Closed, ResolvedPublic

Details

Reference
bz29937

Event Timeline

bzimport raised the priority of this task from to Medium.Nov 21 2014, 11:32 PM
bzimport set Reference to bz29937.
bzimport added a subscriber: Unknown Object (MLST).

PdfHandler passes the PDF file off to GhostScript to render pages to JPEGs... if the fonts are embedded, it shouldn't have a problem unless there's a bug in GS.

Could be a production-level problem (bad version or config of GS).

(In reply to comment #1)

PdfHandler passes the PDF file off to GhostScript to render pages to JPEGs...
if the fonts are embedded, it shouldn't have a problem unless there's a bug in
GS.

Could be a production-level problem (bad version or config of GS).

Hopefully we'll be upgraded to lucid before too long...

Created attachment 8796
gs rendering output of sample file first page on Ubuntu 10.04 Lucid

Here's rendering from my server that's running on Lucid.

Command line:

gs -sDEVICE=jpeg -sOutputFile=- -dFirstPage=1 -dLastPage=1 -r150 -dBATCH -dNOPAUSE -q WM_CZ_2010_Annual_Report_-_Czech.pdf > out.jpg

Appears pretty much identical to the bad output in production.

Attached:

out.jpg (1×1 px, 86 KB)

Which seems to suggest it's probably either a configuration issue, else an upstream bug...

The file rendered correctly for me locally using Ghostscript 9.15:

pasted_file (948×1 px, 224 KB)

…therefore marking as depending on T110849: Upgrade Ghostscript to 9.15 or later.