Description of problem:

It seems Linux Libertine font was updated to 4.7.x-y version. Here comes the changelog:

Changes to version 4.7.0 regular(-) & italic(/) & capitals (C) (20100430)

- All styles: Reported bugs fixed.

- Libertine regular:
	- added missing anchors for i, idieresis, ccaron, scaron
	- added missing characters accent uni0364 and glyph uni1D15
	- completed combining accented chars block 0300-036F
	-ij-Ligature more space between the glyphs, dot of j is slightly higher
	-i,j,... more distinctive dotaccent and higher position
	-r,f,j,y changes in the drop, which became more distinctive 
	- h slightl wider
	- e slightly wider
	- moved [a-z].superior from PUA to dedicated Unicode positions
	- crossed form of W for new Wikipedia-logo

- Biolinum Bold: numerous improvements in following glyphs:
	- a
	- i, j, ä, ö, ü...
	- w, v
	- e
	- f, ß
	- c
	- m, n
	- b, d
	- s
	- kerning faults corrected
	- numbers 0 - 9
	- komma , hyphen -, numbersign #
	- f
	- A-Z slightly thinner verticals
	- R, P better bows


- added small cyrillic m (uni043C) to Biolinum and some other missing cyrillics (not yet all done...)
- Inverted Interrobang (uni2E18) added
- Add Substitution/Transposition brackets 2E02-0A
- macroncmb (uni0304) had a wrong positioned mark anchor and was thus flying much to high
- Problems with Fractions (ID 2665656): without a contextual chaining substitution fractions like 1/7 or 1/10, that don't have a concrete Unicode-posistion, were unavailible. Now solved via contextual chaining substitution.
- bold: ck, ch-ligature's part characters were a bit close. Now some more space between them
- diacritic U+031A was wrongly placed. Is now being positioned above-right. New anchor is called top_rechts
- corrected s_t-ligature in biolinum bold
Ingo Preuss points out thatsome glyphs were wrongly named:
- uni0237 is now called dotlessj (instead of uniF6BE, which was wrongly named 
this way)
- all references to dotlessj now point to uni0237

In the following cases names were corrected, unicode points stayed unchanged
- tcommaaccent (uni0163) is now called tcedilla
- Tcommaaccent (uni0162) is now called Tcedilla
- uni021B is now called tcommaaccent 
- uni021A is now called Tcommaaccent 
- the locl-OpenType-substitution for Romanian wasn't changed, because it already worked correctly
- orthogonal is now called uni221F
- sun is now called uni263C 
- I had to move that .cap-variants of the accents to the PUA, because they covered space where Unicode in between defined new characters. So I systematically moved the .cap-Accents to:
     grave.cap uni0358 -> uniE358 
     acute.cap uni0359 -> uniE359
     circumflex.cap and so forth
     caron.cap
     breve.cap
     hungarumlaut.cap
     space_uni030F.cap
     breveinvertedcmb.cap
     breve.cyrcap 360
     breve.cyr
     dieresis.cap
     hookabovecomb.cap
     dotaccent.cap
(regular, italic)
- new point accent at uni0358
- new Mark for uni0358 in many O/o-characters
- improved support for POJ (Taiwan transcription)
- now kerning for kernclasspair p.sc + a.sc 
- first use of contextual chaining substitution:
   - when a combining accent follows an i, the dotlessi is used to prevent an accent crash
   - when one of the accented or leftwarding glyphs such as (ä, V) follow an f, a short neck version of f (f.alt) is being used for optical reasons
- Added U+033B COMBINING SQUARE BELOW
- there was the "space" in a left kern class (together with dot, etc.). This resulted in too few visible room when a "space" should seperate a word from the now beginning quotation. 
Italic: Correction of schwa