Arabic diacritics in wrong positions

For some mysterious reason, when I install the new font into my system and apply it to an InDesign CC file, the vowel markings are all over the place. I’m sure that the settings in InDesign are correct. The font looks correct in the WOFF test. Can anyone tell me what is happening here? A PDF of the InDesign file, a screenshot of the WOFF test and the FontCreator file are attached. Thanks.
BSE_Higaz_October23.fcp (54.2 KB)
WOFF-Test-Image.jpg
BSE Higaz Test Doc.pdf (48.6 KB)

Randolph, I had a quick look at your files, but I am unable to help as I don’t know about the features used in Arabic fonts.

Others may be able to help you later.

Currently there are a couple of features not fully supported as mentioned here:
Unsupported OpenType Layout Features

Maybe your fonts requires one of those.

Can you please send me a font made with Volt which does work so I can find out what goes wrong?

Thank you for your response. Here is a link to a file which contains both the FontLab, OTF and Volt files.

http://scripts.sil.org/cms/scripts/render_download.php?format=file&media_id=Scheherazade-2.020-developer.zip&filename=Scheherazade-2.020-developer.zip

I’ve taken a quick look at the provided VOLT source file, and it seems FontCreator supports almost everything, but unfortunately it doesn’t understand PROCESS_MARKS (optional field for lookups) and ATTACH_CURSIVE (Cursive Attachment lookups) yet.

PROCESS MARKS defines the lookup’s behavior towards marks. Valid entries are:
ALL - all mark glyphs are processed
NONE - all mark glyphs are skipped by this lookup
Group name - only marks from this group will be processed, other marks are skipped.

NOTE: All marks groups used in this way in lookup headers have to be mutually exclusive. If you need to use this feature, it is advisable that you break the marks down into several separate groups beforehand, and then use these groups in lookup headers. The breaking into groups usually follows the geometrical position of marks (marks above, marks below etc.) and is similar to combining mark classes assigned by Unicode.

Full OpenType support remains high on our to-do list, so I expect Process Marks will be available in the short term.