ZWNJ shows up as space in exported PDF

Bug #327062 reported by Mansour Behabadi
2
Affects Status Importance Assigned to Milestone
Inkscape
Fix Released
Medium
Unassigned

Bug Description

When saving as PDF, Zero-Width Non-Joiner characters show up as a space and *seemingly* hide the character before them (at least that is how it appears).

Zero-Width Non-Joiner (U+200C) is a unicode character often used in languages such as Farsi (Persian) and Hebrew which causes the letter preceding it to take its isolated or final form, and thus not join with the letter following it (http://en.wikipedia.org/wiki/Zero-width_non-joiner).
ZWNJ is a zero width character and thus should not be displayed. Inkscape renders the text with ZWNJ correctly however. The problem only appears when the document is saved to PDF, PNG, EPS formats.

see the screenshot attached for a sample of differences in the Inkscape and PDF version.

OS: ubuntu intrepid ibex
Version: svn revision 20651

Steps to Reproduce:
1. Create a fresh document
2. Using the Text Tool, enter the following text: می‌رفتم و دیده‌ام را غبار غلیظی می‌شست
3. Select "Save As" from File menu.
4. Select Portable Document Format (PDF) and click [Save]
5. Ensure "Convert Texts to Paths" option is *not* ticked
6. Press [OK]
7. Open the saved file and compare with what Inkscape renders on its canvas, you will notice three of the letters are missing.

Revision history for this message
Mansour Behabadi (oxplot) wrote :
su_v (suv-lp)
tags: added: exporting
removed: import-export
Revision history for this message
su_v (suv-lp) wrote :

Is this still an issue with the latest stable release (currently 0.48.1)? If so, could you attach a test file which allows to test this issue on other platforms with different inkscape versions?

Attaching a sample SVG file with the text from the 'steps to reproduce' pasted into Inkscape (using 'DejaVu Sans' and relying on glyph substitution if needed), as well as the PDFs saved with Inkscape 0.48.0 and 0.48.1. Since I'm not familiar with languages such as Farsi (Persian) and Hebrew, it is difficult to assess the correctness of the PDF output (visually it seems correct to me).

The PDF file created with Inkscape 0.48 apparently misses some characters (some appear displaced, with latin glyphs?), whereas the PDF file created with Inkscape 0.48.1 renders identically in Apple's PDF viewer (Preview) and evince 2.30.3

When saving as PDF with Inkscape 0.48.1, these warnings are output to the console:
** Message: Invalid glyph found, continuing...
** Message: Invalid glyph found, continuing...
** Message: Invalid glyph found, continuing...
** Message: Invalid glyph found, continuing...
** Message: Invalid glyph found, continuing...
** Message: Invalid glyph found, continuing...

(tested with 0.48.0, 0.48.1 and r10325 on Mac OS X 10.5.8 (i386))

Revision history for this message
su_v (suv-lp) wrote :
Revision history for this message
su_v (suv-lp) wrote :
Revision history for this message
su_v (suv-lp) wrote :

Feedback about the state of the reported issue with the latest stable release (0.48.1) would be appreciated.

Changed in inkscape:
importance: Undecided → Medium
status: New → Incomplete
Revision history for this message
Mansour Behabadi (oxplot) wrote :

Based on your 4.81 version PDF, as well as my own testing on Inkscape 0.48.1 r9760, I can confirm that this problem no longer exists. I am happy for this bug to be closed.

Revision history for this message
su_v (suv-lp) wrote :

Thank you for your feedback. Closing this report as "Fix Released" in 0.48.1 - please reopen it if the problem reoccurs with the latest stable release.

Changed in inkscape:
milestone: none → 0.48.1
status: Incomplete → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.