[poppler] auto switching of TrueType hinter for Type42/Type11 font in PDF

[mpsuzuki at hiroshima-u.ac.jp]

Hi,

I put a Subject for the discussion about the question:
which library should work for this problem? freetype2?
poppler? or both of them?

On Wed, 25 Aug 2010 19:35:34 +0100
Albert Astals Cid <aacid at kde.org> wrote:
>A Dimecres, 25 d'agost de 2010, mpsuzuki at hiroshima-u.ac.jp va escriure:
>> >So, how can I fix poppler on my computer?
>> >Or do I need to wait until the next release of poppler?
>> 
>> About your own solution, there are several ways.
>> 
>> 1) Modify application personally.
>> 
>> This is Cobra suggestion for Qt4 application. Although
>> poppler-qt4 has a switch to enable hinting in all fonts
>> in a document (disabled by default), Okular does not
>> touch the switch. So the insertion of the enabling of
>> the switch will improve the quality of Okular result
>> for PDF including MingLiU.
>
>That is giving the user the responsability of making things work, as a last 
>resort is acceptable but not the best solution in my opinion. In my opinion 
>things "should work" without user interaction, and that means that either 
>FreeType or poppler need to make "things work" automatically,

Thank you for comment. I agree. The 3 "solutions" are
proposed as the personal solution without waiting for
other libraries upgrades. In fact, at the end of the post,
I wrote as:
>> Therefore, I'm trying to propose a fix that can be
>> accepted by the maintainers officially.

In next,

>and in my 
>opinion FreeType makes more sense just for the reason that is lower in the 
>stack and improvememnts there will mean improvements everywhere, not just in 
>poppelr.

again I'm glad to hear your opinion. As I've written in
previous post:

  http://lists.freedesktop.org/archives/poppler/2010-August/006286.html

it is difficult for FreeType2 to determine hinting is
required /or not for an embedded True42/Type11 without
"name" table, because FreeType2 does not receive reliable
identifier. The additional information from PDF parser
layer is expected. The auto switching of the hinter for
the tricky font is generic problem, as you wrote, but
I think the handling of TrueType without of "name" table
issue is PDF specific issue. How do you think of?

Regards,
mpsuzuki

P.S.
Just I've started the experiment to identify the tricky
font without "name" table, but it's just experiment.