Hi Tim - re: Unicode - right now MoI is using the simplified UCS-2 encoding in this case where it strictly uses 2 bytes per character which doesn't allow for the surrogate pair mechanism needed to handle planes above 0.

Normally this isn't much of an issue since plane 0 covers all the stuff in use in current living languages... The main area where it doesn't work are for glyphs used in ancient languages like you're working with here.

I think I should be able to fix it up, but it would be helpful if I could have a sample to test with instead of trying to work on it blindly.

Do you know of a particular font that I could download and install and some character codes that I could use to test with over here that don't currently work so I can reproduce the problem over here and test the fix?

- Michael

Hi Michael,

I'm at work just now but will try that workaround when I get home. (yes, I know MoI can be carried on a key but my work computer is running windows 2000 and has a graphics card that isn't up to it :) )

For the fonts you could use any of the free fonts available at http://users.teilar.gr/~g1951d/
I've tried out characters from 3 or 4 of them when I was testing/playing about.
The one in my example is the Aegean font using a few characters from Linear B including unicode # u+10024 and u+10025
The Unicode Symbols font might be good for testing also since it includes a bunch of technical and mathematical symbols that someone might actually want to use some time.
This isn't any huge issue for me, just something I ran across when working on a sci-fi art project in Vue. Vue isn't able to handle any of these characters either. My new desktop is powerful enough I can work in both Vue and MoI at the same time, hopping into MoI whenever I want to create some unique object for a scene. It makes for a nice creative workflow.

Hi Tim, thanks for the link.

I've looked into this today and unfortunately it appears that there are limitations in Windows which prevent this from working.

MoI calls a Windows function called GetGlyphOutline to retrieve the geometry for a given character glyph.

Previously, MoI was not decoding "surrogate pairs" properly so it would see Unicode characters above U+FFFF as 2 garbage characters instead of as a single character value.

I have now fixed this part of MoI up for the next v2 beta so that it can correctly decode character values above U+FFFF now in the Text command. However, when MoI calls the Windows function GetGlyphOutline with a character greater than FFFF, the Windows function ignores the high bits, it only seems to use the low 16 bits.

So for example when MoI requests the geometry for character U+10024 , instead Windows decides to give the geometry for character U+0024 instead.

I've got it all set up now so it passes the correct value to Windows so hopefully in the future if Microsoft can correct this problem it will then start to work. But until then there doesn't seem to be anything that I can do about it, short of developing a completely custom font parser that hits the font data directly rather than using the Windows API call. Unfortunately it would take a significant amount of work to accomplish that. There are actually some libraries out there that could help with this part but even incorporating a library can still take a fair amount of time.

If you have any other programs out there that are able to extract geometry from Unicode characters above U+FFFF, please let me know and maybe I can see what they are doing to accomplish that. But my guess is that other programs will likely have the same issue as MoI.

Other parts of Windows are able to understand the special characters such as the font rasterizer, but it seems that this specific GetGlyphOutline function with the GGO_NATIVE flag unfortunately was not updated along with the rest of Windows when Unicode was expanded to have these special new characters. The original Unicode 1.0 specification did not include these characters and only had characters with values up to U+FFFF.

- Michael

Thanks a lot for looking into it Michael. I wouldn't even have attempted a ticket on this with Vue or most other programs. MoI is just so good it gets held to a higher standard. ;)
I appreciate your taking the time on this. It certainly isn't important enough to take the time for coding a special parser or anything. (The layers and colors crowd would have my scalp :D )

You bet ! :)
eric

You have a very simple thing to make :)
Just type your text with any True Type Police in the Window caracters' table and copy paste Inside Moi Text !!!
so try with the "Analectica" from the site above
That works like a charm :)
Have fun hieroglyphics :D
Edit : Alas seems not work for extra Unicode like U + 2CB7 :(

Ok more subtile !
Type your Unicode at the end of the follow line here 2CB7
http://unicode.org/cldr/utility/character.jsp?a=2CB7
goto the page result copy / past the character inside Moi et voilà :)
for another just modify the end address in your navigator :)
Only default it's letter by letter for extra Unicode! :)

Hi Pilou, that should work fine for all Unicode characters in the range of U+0000 to U+FFFF, which includes characters for all current languages in use around the globe, and also quite a lot of symbols are in there too - this is for the Unicode 1.0 standard.

However, in later versions of the Unicode standard they ran out of space in that range and have produced new characters for some specialized purposes like support for archaic languages, etc...

It is these new characters that will not work, due to a limitation in Windows which I described above. But they tend to be not used very often, since for example people don't write very often in Cuneiform these days...

Windows does support these extended characters in other parts like for drawing the text on the screen with the Windows drawing methods. Unfortunately it does not support it in the particular function that MoI needs to use which involves extracting the bezier path information of the characters.

- Michael

You right!
We can see the result for the Cuneiform Font Cuneiform Akkadian, Unicode 12219
http://unicode.org/cldr/utility/character.jsp?a=12219
but the copy past don't work! :(