Hello there, I am developing a CMS Software using MSHTML for about three years, and I get this behavior since I started using MSHTML. Please excuse that I write such a big text but I think I can't express this in a shorter way. The text pages are solely using the TiresiasScreenfont, the user has no option to change the font (this is mainly due to restrictions on the target platform). Now sometimes, the font changes to Times within a line and back to Tiresias in the next line. I have also seen lines of text where the upper half of the line is in times and the lower half in Tiresias, which really looks terribly. The only solution is to cut the complete content of the page, paste it into notepad and copy and paste it back again so that default formatting is applied. The behavior occurs somehow randomly but has always something to do with the text - e.g. overlapping font tags in connection with German umlauts, like <font color="#CCCCCC>Hauptmenü<font color="#EEEEEE"> / </font> Information</font>. (Here, the Renderer seems to switch to Times at the word Information) I use a derivate of CHtmlEditView with two DocHoistUIHandler interfaces which I use to prevent the user from making not allowed changes (i.e. inserting more than one picture per page). I assume that the problem is either with the MSHTML editing control or with the Tiresias font. Has anybody experienced something like that? Here's a screenshot of the editing area: http://www.leoloewe.net/cp-screenshot.gif[^] Would be nice if somebody with more experience would answer my post... ;) Thanks in advance, Leo Thou Shalt Process Messages If Thee Wishes Thine UI To Behave Properly (Gary R. Wheeler) -- modified at 2:52 Thursday 20th July, 2006 P.S. As the behavior doesn't occur with any other font, I assume that the Tiresias is responsible for that. If somebody read this, please don't use the quite expensive Tiresias ($17000 each) fonts in your projects if you plan using MSHTML.

I experienced a strange behavior with overlapping tags. I use the IMarkupServices interface with IMarkupPointer to parse a document and display the "class" attribute of the currently selected <font> tag, if any (just like Winword displays the style of the current selection). When I parse the following code (after selecting the text portion marked green in the view), strange things happen: some text<font class="g3">Normal<font class="g4">Big</font>Normal</font>some other text The IMarkupPointer pHtmlStart starts at the right point, but when it moves on after looping once and successfully retrieving the "class" attribute, pHtmlStart->CurrentScope returns a NULL element and seems to have gotten RIGHT of the pHtmlEnd pointer. Due to this behavior, I cannot retrieve any other elements in the selection, thus I am not able to indicate "there are elements with different CLASS attributes in the selection.". Has anybody experienced something like this? Or is there another solution to the problem? Anybody knows how FrontPage handles this? Or am I doing something totally wrong? If anybody is interested I will post some code here to evaluate the behavior. Thanks for reading. -- modified at 3:18 Friday 7th July, 2006

I recommend Rob Manderson's article: Capture an HTML document as Image[^]