Apache OpenOffice (AOO) Bugzilla – Issue 78848
Saving as HTML misplaces Hebrew diacritics
Last modified: 2017-05-20 11:29:24 UTC
Hebrew documents sometimes have diacritic marks for vowels (niqqud). These marks are positioned close to the space of the consonant that they mark, sometimes above, sometimes below, sometimes within. They are displayed correctly in Writer. However, if I export to HTML, and view the exported file in a browser, the vowels are treated as consonants, and have their own space. This is incorrect. See screenshots.
Created attachment 46230 [details] The document as it appears in Writer (correct)
Created attachment 46231 [details] The HTML exported document (incorrect)
Reassigned to ES.
Created attachment 46234 [details] Sample document used to make screenshots
The browser screenshot that I provided was from opening the file on Firefox. When I opened the file in IE, the text looked OK.
Comment from Jonathan Ben Avraham: The problem is probably one of the following: 1. No diacritical heuristics The browser is not using the Uniscribe rendering engine and the rendering engine it does use does not support OpenType and does not have it's own Hebrew diacritical heuristics (as does Uniscribe). 2. No support in font The browser uses FreeType II or better rendering engine with OpenType support but does not use an OpenType Hebrew font with diacritic support. If the browser uses Uniscribe with a font that does not have OpenType GPOS/GSUB diacritic support, then the Uniscribe engine has heuristics that can render readable pointed Hebrew text.
@HDU: please have a look if it is something you can fix. Else, if it is a pure HTML export missing feature close it as WONTFIX. The HTML export developpement is stalled due to a lack of ressources.
I agree that it is either a browser problem or a OOo html-export problem. I'm sure that recent browsers have no problem with this, so what needs to be done is analyze the exported html and improve this export to help the browser. This looks like an interesting task for an OOo html-export expert => reassigning @FME: thanks for volunteering
fme->ayaniger: We first have to clarify if this is a Writer or browser issue. As you already mentioned, IE shows the file correctly. I expect the bigger CTL competence to be on your side. So please have a look.
With OpenOffice.org 3.2.1 OOO320m19 (Build:9505) and Firefox 3.6, the diacritics looks OK. The bug can be closed as worksforme.
Reset assigne to the default "issues@openoffice.apache.org".