[poppler] Wrong hinting

[Cedric Sodhi]

Thank you very much for this explanation. The issue has bothered me for
a long time, I hope this file can be closed for good now.

On Mon, Feb 14, 2011 at 10:48:28PM +0900, mpsuzuki at hiroshima-u.ac.jp wrote:
> Hi,
> 
> On Mon, 14 Feb 2011 14:25:12 +0100
> Cedric Sodhi <manday at gmx.net> wrote:
> 
> >I'm pleased to announce that a update of several packages has resolved
> >this issue, among which Freetype was updated to 4.4 (I had wrongly
> >stated that I already had 4.4 when I checked it while the update was
> >running).
> 
> Thank you for notice. By statically built binary with the
> latest FreeType2, the issue disappeard in my side too.
> I guess, freetype-2.4.3 was the root of this issue.
> 
> Before freetype-2.4.2, most TTFs are rendered without
> hinting, when no-hint flag is given.
> 
> In freetype-2.4.3, nameless TTF is rendered with hinting,
> even when no-hint flag is given. This change was introduced
> for tricky Chinese TTF which break their shapes unreadably
> if rendered without hinter. Originally, such font was
> detected by the font family name, but embedded TTF in PDF
> often lacks its family name. Thus, there was a proposal
> to enable hinting for nameless TTF in PDF. It was adopted,
> so freetype-2.4.3 enables the hinting for nameless font.
> 
> Afterwards, some PDF browser users complained that still
> FreeType2 hinter makes worse results for some Latin fonts
> and "nameless font" is too loose condition to enable hinting
> forcibly. So, I've introduced another way to identify the
> tricky Chinese TTF even if the font is nameless. This new
> detector is included in freetype-2.4.4, so, this issue
> mignt disappear.
> 
> >Thanks for your help so far. If I can do something to help you pin this
> >down for the sake of developing freetype, please do so! I also would
> >like to know what is (was, lucklily) the cause of that.
> >
> >Cedric
> >
> >PS: I'm not hundred percent sure that the update of freetype did the
> >job, If possible, I will dowgrade FT later and see whether the propblem
> >reappears.