No matter what you do to keep the formatting you still need the words in a plain text to run through your spell checker. Using the text property saves all the hassle of removing the html formatting. Thanks Adobe! Things like "<b>appli</b>cation" come back correctly as one word.
When I get the plain text i run this regex "/\b\w+/gi;" to parse out all the words. Removes punctuations and such, parsing just on spaces won't do. It's not 100% perfect but can finetune later. Once I know a word is not spelled correctly against my master word list and have the user's corrected spelling, i go find it in the htmlText and weave it in. The only thing that does NOT work really well is when a word is formatted differently in itself such as "<b>appli</b>cation". I currently lose that the first 5 letters are bolded... oh well. Can fix later. What i've found most difficult is the UI to highlight the word so a user has context. For now i'm just using the setSelection since i'm spell checking on TextAreas only. Not sure how you are doing suggestive words but havn't firgured that out yet, no algorithm. I'm loading a 120,000 word list into an array collection :)! I'm quite impressed how fast it is, loads in milleseconds when i split my text file on \r\n. Using the build in contains to validate each word. It's faster than calling any other web service. Would like to know what you are doing for suggestive words. --Scott --- In [email protected], "jmfillman" <[EMAIL PROTECTED]> wrote: > > Ultimately I'm wanting to find each word to check the spelling, while > maintaining the HTML formatting of the TextArea. Using the text would > be far easier, but I loose the formatting. >
