Auto capitalisation

Works ok, but works for stuff like abbreviations and such too.

I implemented auto capitalization a long time ago (about 1.5 years) using an override an the following procedure.

TWPCustomRtfEdit.DoAfterCompleteWordEvent

There is also a property AfterCompleteWordEvent that can be used.

The only problem is that DoAfterCompleteWordEvent isn't triggered when starting a new paragraph, so i had to trigger it myself.

I remember BeforeInitializePar not being sufficient, becuase it didn't trigger immediately and also because i liked it to work like Ms Word where you can override the auto-capitalization by retyping the same word in lowercase, but this does require some more work to implement.

[/code]

Zitat von RoyLambert

Thanks I tried searching on autocaptialise, never thought of putting a hyphen in <g>

I only knew it was there because I asked the same thing a while back :)

The code is somewhat integrated with the rest of our application. And i'm not allowed to share our source code, but i can give some code segments.

I made this code when we still used WpTools 3 and converted it when we upgraded to WpTools 5, so it's not by any means optimized for WpTools 5. I can't even guarantee that everything still works, since i haven't maintained this code through many releases of WpTools.

TCsTreeDoc = class(TWPRichText)


{i can't give you the entire implementation of this method, but i'm sure there are better ways to retrieve the word at a certain position and the position of the first character in the word, then what i implemented here.}


function GetWordFromPos(pvPos : integer; var pvStartPositionRecord : TPositionRecord) : string; 


procedure TCsTreeDoc.CheckCapitalAtCp;
var
  lvNewWord : string;
  i : integer;
  lvStartPosRecord : TPositionRecord;
begin
  lvnewWord := GetWordFromPos(CpPosition, lvStartPosRecord) ;


  if lvNewWord = ''
  then exit;


  for i := 1 to length(lvNewWord) do
  begin
    if not(lvNewWord[i] in ['a'..'z'])
    then exit;
  end;


  i := lvStartPosRecord.pos_in_par;
  while (ActivePar.IsWordDelimiter(i - 1)) and (i <0>= 0)
  do dec(i);


  if (ActivePar.ANSIChr[i - 1] in ['.','!','?'] ) or (i = 0)  then


  if Uppercase(FWordBeforeComplete) <> Uppercase(lvNewWord) then
  begin
    FWordBeforeComplete := lvNewWord;
    ActivePar.CharItem[lvStartPosRecord.pos] := WideChar(Integer(ActivePar.CharItem[lvStartPosRecord.pos]) - 32);
    ActivePar.Refresh;
  end;
end;


procedure TCsTreeDoc.DoAfterCompleteWordEvent(var key: char);
begin
  inherited;
  if csAutoCapitals in CsAutoOptions
  then CheckCapitalAtCp;
  DelayedReformat; //this may or may not be necessary anymore
end;


procedure TCsTreeDoc.KeyDown(var Key: Word; Shift: TShiftState);
begin
      lvCurrentWord := GetWordFromPos(CPPosition, lvPosRecord);
      if FWordChangedOnPos <> lvPosRecord.pos then
      begin
        FWordChangedOnPos := lvPosRecord.pos;
        FWordBeforeComplete := lvCurrentWord;
      end;
    case Key of
      VK_RETURN: DoAfterCompleteWordEvent(c);
    end;
end;

[/code]

Zitat von wpsupport

Hi,

This following code is a new approach. It seems to work pretty nice and allows it to undo the auto change - something I find important. It handles abbreviations! Any suggestions are welcome.Regards,

Julian Ziersch

for me, this is doing p.M. for abbreviations. (p.M. hour) For three letters it does 'a.B.c'. The word following the abbreviation is good though.

Is this correct?

// Check for abbreviation (simplified - only check a..z
and ((par.IsSpace(i-1) or (par.CharItem[i]<or>='a') and (par.CharItem[i-1]<='z') and par.IsSpace(i+1) )) )
then

Ooooh. That works nicely. Thanks!

Here's a new wrinkle.

When the sentence ends with a number, the next word is not capitalized.
ie ' There was 123. but blah blah'

Also, it is capitalizing words that follow the start of a sentence where the first word is a single char.

like

I Am here.
A Day at work.