[okular] [Bug 367959] New: Okular mangles some characters (non embedded fonts)

[Sergio via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=367959

Bug ID: 367959
Summary: Okular mangles some characters (non embedded fonts)
Product: okular
Version: 0.24.0
Platform: Kubuntu Packages
OS: Linux
Status: UNCONFIRMED
Severity: normal
Priority: NOR
Component: PDF backend
Assignee: okular-***@kde.org
Reporter: ***@gmail.com

In some occasions, Okular mangles some chars that are reproduced correctly by
other PDF readers (e.g. acroread, masterpdfeditor) on the very same platform.
Inkscape also imports the PDF correctly. Might be a poppler problem since xpdf
has the same issue. The problem seems to be related to substitution rules for
fonts that are not embedded in the PDF document nor available on the platform
with exactly the same name.

Reproducible: Always

[Sergio via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=367959

--- Comment #1 from Sergio <***@gmail.com> ---
Created attachment 100839
--> https://bugs.kde.org/attachment.cgi?id=100839&action=edit
Demo file

A demo file to reproduce the problem

[Sergio via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=367959

--- Comment #2 from Sergio <***@gmail.com> ---
Created attachment 100840
--> https://bugs.kde.org/attachment.cgi?id=100840&action=edit
How okular shows the demo file

[Sergio via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=367959

--- Comment #3 from Sergio <***@gmail.com> ---
Created attachment 100841
--> https://bugs.kde.org/attachment.cgi?id=100841&action=edit
How other pdf readers show the file on the same platform

[Oliver Sander via KDE Bugzilla]

https://bugs.kde.org/show_bug.cgi?id=367959

Oliver Sander <***@tu-dresden.de> changed:

What |Removed |Added
----------------------------------------------------------------------------
CC| |***@tu-dresden.de

--- Comment #4 from Oliver Sander <***@tu-dresden.de> ---
Confirmed on

Qt: 4.8.7
KDE: 4.14.23
Okular: 0.25.0
poppler 0.44.0
Debian Testing

[Sergio]

https://bugs.kde.org/show_bug.cgi?id=367959

--- Comment #5 from Sergio <***@gmail.com> ---
Still seen in kubuntu 17.04 with the KDE PPA, that is

Okular 1.0.3
Framework 5.35
QT 5.7.1
Poppler 0.48.0

Please, mark as confirmed

[Christoph Feck]

https://bugs.kde.org/show_bug.cgi?id=367959

--- Comment #6 from Christoph Feck <***@kde.org> ---
Unless a developer confirms if the issue is caused by Okular code, poppler
code, or just because of a font or configuration issue, it does not make sense
to change the status.

[Jaime Torres]

https://bugs.kde.org/show_bug.cgi?id=367959

--- Comment #7 from Jaime Torres <***@gmail.com> ---
Created attachment 116379
--> https://bugs.kde.org/attachment.cgi?id=116379&action=edit
The demo file rendered by okular (with arial font installed)

[Sergio]

https://bugs.kde.org/show_bug.cgi?id=367959

--- Comment #8 from Sergio <***@gmail.com> ---
Can you please specify better on which platform you made the last screenshot?

I am on kubuntu 18.04 and I still see the issue.

The demo file is clearly *not embedding* its fonts and okular has to pick some
fonts that it finds on the system for its rendering, which is the factor
triggering the font mangling. However, the requested font is
`TimesNewRomanPSMT`, so the fact that you have an *arial* font installed on the
system should not make any difference and if it does, this looks like one more
weirdness. Nonetheless, the font that is being substituted for the required
`TimesNewRomanPSMT` is `dejavusans.ttf` on my system, which is a sans font,
just like your arial so it looks like the system font machinery does not
correctly match TimesNewRomanPSMT to a Times New Roman font.

The document uses `TimesNewRomandPSMT` with two different encodings, one is
`WinAnsi` and the other one is `Identity-H` with some internal mapping to
unicode.

It looks like it is this Identity-H encoding to be troublesome to Okular.
Interestingly, also MasterPDFEditor recently mangles the fonts in the same way
as okular.