The "½" character does not print correctly.

[Nigel Broad]

Binary package hint: evince

On page 9 of the attached pdf file the "½" character in the formula about half way down the page does not print correctly: a horizontal line is printed in place of the correct character. The effect is apparent in the print preview too. This is in Ubuntu 9.10. Evince version 2.28.1-0ubuntu1.2.

[Nigel Broad (n-broad)]

I tried the same pdf file using Okular and it prints correctly.

[Dennis Sheil (dennis-sheil)]

I have the same result, also Ubuntu 9.10 and evince 2.28.1-0ubuntu1.2 (and poppler 0.12.0-0ubuntu2.1).

For evince version 2.26.2 and poppler 0.10.5, this does work. For the version you mentioned, it has stopped working. I will try to track down which commit broke this. I suspect it was a poppler commit somewhere which broke it.

[madbiologist (me-again)]

Confirmed on Lucid Lynx 10.04.

Uname: Linux 2.6.32-10-generic i686
Package versions:
evince 2.29.5-0ubuntu1
poppler 0.12.2-2.1ubuntu3

[Dennis Sheil (dennis-sheil)]

This used to not be a problem with poppler, but after a code commit on November 1, 2008, it stopped working. The commit that broke this dealt with Type 3 fonts, so I suspect the 1/2 fraction is a type 3 font and that commit broke it.

I've reported this upstream to poppler as they are the ones who would create a patch for this.

[madbiologist (me-again)]

Ever since poppler commit a3edfa30680864b95a5196c5619846de42980857 - "Implement Type 3 fonts in cairo backend using cairo user-fonts" on November 1, 2008, document viewers which use poppler (such as Evince and Epdfview) are unable to print (or print preview) a complex glyph created from a type 3 font, namely the fraction 1/2. The fraction is displayed correctly on screen but is not printed (or print previewed) correctly.

Please see fdo bug #26143 for more info and a test case.

[Adrian Johnson (ajohnson-redneon)]

Fixed in http://cgit.freedesktop.org/cairo/commit/?id=b7550bf0f1cafb5cc5df13da50b0cec80313038f

[madbiologist (me-again)]

Wow - that was fast!

Thanks mate.

I'll update the Ubuntu launchpad bug report.

Happy Australia Day to you.

[madbiologist (me-again)]

Fixed upstream in cairo git.

See http://bugs.freedesktop.org/show_bug.cgi?id=26186 for details.

[madbiologist (me-again)]

Is anyone any good at compiling from git?

[Adrian Johnson (ajohnson-redneon)]

Created an attachment (id=32802)
Patch for 1.8.8

The fix introduced a regression. I've fixed it now with 180aea76

I'm attaching a single combined patch for 1.8.8 since the two patches in master don't apply cleanly to 1.8.8.

[Sebastien Bacher (seb128)]

The corresponding commit is http://cgit.freedesktop.org/cairo/commit/?id=b7550bf0f1cafb5cc5df13da50b0cec80313038f

[Sebastien Bacher (seb128)]

the change backported from git to 1.8.8 doesn't seem to fix the issue

[madbiologist (me-again)]

Sebastien: Thanks for testing (an thanks for all the work you do in all the other Launchpad bugs). Apparently the original fix introduced a regression. I'ts been fixed now in a new patch for 1.8.8, available at http://bugs.freedesktop.org/attachment.cgi?id=32802

See http://bugs.freedesktop.org/show_bug.cgi?id=26186 for more details.

Are you able to try this one? If you are able to put it in a PPA (preferably a Lucid PPA, but a Karmic PPA is OK) I will try it too.

[Adrian Johnson (ajohnson-redneon)]

Sebastien: If you print page 9 to PS without the fix, open the PS file in a text editor and search for "Type 3", about 9 lines down you should see:

BT
55.728 0 0 -51.6 0.00343792 0.0108895 Tm
/f-0-0 1 Tf
[<11>500<26>]TJ
ET

With the fix applied the output is:

BT
0.556723 0 0 0.515484 0.0481308 -0.152452 Tm
/f-0-0 1 Tf
<11>Tj
0 0.8961 Td
<26>Tj
ET

In the PS file you sent me it contains:

BT
-0.556723 0 0 -0.515484 0.0481308 -0.152452 Tm
/f-0-0 1 Tf
<11>Tj
0 -0.8961 Td
<26>Tj
ET

Note the '-' sign in front of the 1st and 4th number and in front of the number before "Td". My guess is something in the patch (probably the "invert_y_axis" part) has not been applied correctly in your version.

[madbiologist (me-again)]

Still occurring on Ubuntu 10.10 "Maverick Meerkat" alpha2, updated to the latest packages.

Uname: Linux 2.6.35-6-generic i686
Packages: evince 2.30.3-1ubuntu2
                  poppler 0.14.1-0ubuntu1
                  libcairo2 1.9.12-1build1

[madbiologist (me-again)]

This is fixed on Ubuntu 14.04 "Trusty Tahr".

evince 3.10.3-0ubuntu10
poppler 0.24.5-2ubuntu4

[madbiologist (me-again)]

Oops, I should have added:

libcairo2 1.13.0~20140204-0ubuntu1