Apache OpenOffice (AOO) Bugzilla – Issue 82950
hebrew niqqud vowel: hiriq below yod too far up
Last modified: 2017-08-20 18:49:31 UTC
The hiriq vowel is displayed too far up when written below a yod consonant, see attachments. The same shows up with both fonts (Ezra SIL and SBL Hebrew) that I have installed. I use KUbuntu 7.10 (KDE) with a biblical Hebrew input method (kmfl). On my system, I found that only Qt4 supports Unicode (in whatever ways) far enough to display biblical Hebrew properly.
Created attachment 49129 [details] The document displaying the text "Israel Jerusalem" in two fonts
Created attachment 49130 [details] Screenshot of OOo displaying incorrectly; Qt4 displaying correctly (see rightmost character)
Reassigned to SBA.
For those who don't have background in Hebrew computational linguistics, here's the problem as I understand it: Biblical Hebrew uses self-standing consonant letters (Unicode U+05D0-U+05EA) which are usually coupled with lots of different combining diacritics (U+0591-U+05C7). Here we have U+05D9 (yod) coupled with U+05B4 (hiriq). In Hebrew typography, the latter is usually placed just below the base line. In this case, however, there's a problem, since U+05D9 + U+05B4 corresponds to a precomposed Unicode character U+FB1D, which often has its own glyph in many fonts. The form of this glyph is not suitable for Biblical Hebrew but just for Yiddish, where it is sort of used as a consonant letter of its own. So OOo probably shows the precomposed glyph, while Qt4 combines them on the fly. I know that Ezra SIL, at least, is specifically designed for Biblical Hebrew, so there is probably a way to get the font not to show the precomposed Yiddish glyph, by using some OpenType / Graphite / etc. feature. It might even be something depending on the document language. Those who are more familiar with the different font technologies can probably sort that one out.
An example of when the diacritics should not remain below the baseline in Biblical Hebrew (rendered correctly both in OOo 2.3 and Qt 4 with Ezra SIL): ךְ (U+05DA HEBREW LETTER FINAL KAF and U+05B0 HEBREW POINT SHEVA) should have the sheva centered below the *upper* line of the final kaf. (Afaik - the mentioned implementations look correct in my view at least, but I didn't do any thorough research on this.)
Any progress? If I got to fix this, where would I start looking at the OpenOffice code?
kaplan -> sba: I can confirm the bug exists in 2.4.0 (tested on Debian GNU/Linux). Please change the status to NEW.
confirming (OOo 2.4.1)
Is there any suggestion for a temporary workaround for this issue?
It’s the same thing in latin characters: if you want to display a t with a caron (inverted circumflex) on top, all the fonts I’ve looked at give you the precomposed ť (U+00165) instead.