Hello Xichen.

I think you have to do this by yourself. The library has no help in this case. But the library also won't block this. To combine subscript and superscript you need to use three different fonts. The super and subscript fonts needs to be an bit smaller and an offset done with baselineoffset. To draw these texts you can use easily the blockmode. But in the blockmode you can't controll any of the text. So i think the best way is to use the singlelinemode and construct the text by yourself. You have to store informations about the text. The user typed an word and the he switched to subscript and typed an other word. You have to store the informations what the user has done. For example your structs (or whatever you want) contains: (normal) 'Sample'#32 (subscript) 'subscript'. If you wants to draw your text you activate the Font and draw the text. For the next word you hav to query the length of the part and add them to your offset.

To draw an carret you have to save the position of the carret in chars inside your edit. Means the carret is after the char number 12. To calculate the position of the char you have to query the length of the whole parts before the carret. If the carret is inside an part you have two different ways. Create an text they only contains all chars until the carret position and query the length of them. Or you query the advance of every char and add them to you offset. Avoids text creation but results in more library calls. At the calculated position you can draw your carret. Thats "all".

To support mouseclicks you have to do somethink like before. Test in which textblock the user clicked and then test which char is below. But for this quering every single char is better than creating many texts they only one char longer that the previous. You should consider this in your plans for the future of your edit.

Only chance to optimize is to cache calculated values of the blocks/chars. To avoid quering always the same values from the library. It should be faster if you store this directly in your structs.

Greetings

Dear Steffen,

I thought I have thanked for your helpful comments but apparently I didn't!! I am sincerely sorry!!

Best wishes,

Xichen Li

Yes the function TtsRendererOpenGL.DrawChar draws one single char. This function gets called from DrawLine. But there is no way to get or collect the vertex data in the current structure. But it's not so easy as you think. The chars won't stored in one texture. If you have luck your chars could split over two or more textures. Because the texturespace manager stores the chars whereever enough space is. If it's not enough he will create an new empty texture and use this. So if you plan to use vertexarrays you have to sort the chars of used textures.

In the future i have the plan to support vertext buffer objects. But not in that way you want to do. Only to speed up line output or complete blocks. But not to combine completly different calls.

Maybe you can speedup your code if you use vertex arrays or vertex buffer object for your circles to draw them once and than draw all texts at once. But this won't work if you have overlapping circles. If you want to speed up this much more it's could be better to use an offline font generator. Only create the chars you need in one texture and than you have to draw textured quads. So you have the full control to use vertex arrays for your drawing and so on. The TextSuite was more designed to draw complex text blocks with different fonts/sizes without knowing much from the fonts and the design of fonts.