Text problem in word when using COM

[kcvinu]

Hi all,

I have made a simple script to monitor the text which user types in a word document. This script is working like this.

1. If word is opened then get the object of word.application.

2. Entering to a while loop

inside that loop (sleep = 50)

   1. check for current cursor position and get the word behind the cursor.

    2. check if that word is in a dictionary and if yes, paste the key of the dictionary and again monitor another word

  So, this is the working style of that script. But when i run this script, everything is ok but the text in word looks very strange. I am attaching here an image of the difference. This problem only occure in a specific template. Not in a blank template. I have tested the same script idea with vb.net. there too i faced the problem. See the image. I need the Actual letter. But now i am getting the current result. 

How to fix this.

[Melba23]

kcvinu,

I have made a simple script to monitor the text which user types in a word document

That sounds very much as if it comes under the keylogger prohibition - can you persuade me that it does not?

M23

[kcvinu]

@Melba23 Sure. It is nothing but a text expander. I am TV program translator by proffesssion. I need to type thousands of words per day. So i need use a good tex expander. I have Phrase Express but it won't support my native language. So i learned autoit and made one for me.

[Melba23]

kcvinu,

I have no problem with the fact that the script is not a keylogger by design, nor that you have such a script for personal use. However, please be careful with any code that you post on the forum - it sounds as if the full script does rather more than "check for a few keys being pressed" which is the touchstone for acceptability.

M23

[kcvinu]

@Melba23  Sure. Infact, i am learning autoit for my own purpose. I don't want to make harm for anybody. 

[jchd]

kcvinu,

I understand your issue, but why did you post a bitmap and not text? Doing so doesn't allow to grab characters as you want them.

Anyway, I don't know malayalam enough to guide you to a final resolution but consider that both ൺ (MALAYALAM LETTER CHILLU NN, codepoint 0x00D7A) and ട (MALAYALAM LETTER TTA, codepoint 0x00D1F) are two distinct letters. Only a malayalam-aware text processor will be able to detect that they can be represented by a ligature as in your "actual letter" example [I guess it's a ligature but not knowing enough about the script I may be wrong]. AFAIK there is no Unicode codepoint attributed to this ligature and it can only be detected and represented by a malayalam-aware font rendering code.

I'm certain your code isn't intended to do anything bad and I also understand Melba concern. Can you please copy and paste the "actual letter" in text form (just like I did). Leave code away for now.

I advise you read from this page (but reading the entire site would help even more).

[kcvinu]

@jchd Thanks for the reply. This is the actual letter. " ണ്ട "

[jchd]

This is where you're confused: the text you just posted is not a letter, to wit:

#include <Array.au3>

Local $s = "ണ്ട"
Local $a = StringToASCIIArray($s)
_ArrayDisplay($a)

You can as well copy the first character or the last one with the mouse on the post. This is a string of three Unicode codepoints, namely 0x00D23, 0x00D4D and 0x00D1F. That is, in this order :

MALAYALAM LETTER NNA
MALAYALAM SIGN VIRAMA
MALAYALAM LETTER TTA

The actual rendering engine (depends on the font you use) decides to group these three glyphs into a new one which doesn't have a unique Unicode codepoint. The link I gave above explains what happens, even if there is little malayalam example, you'll get the idea.

Also as MS Word does a large number of things in your back, I'm uncertain you can succeed in making it do exactly what you want. Maybe simply by creating a user ligature in the menus.

[kcvinu]

@jchd Thank you for the reply. I have visited the link. I think you have pointed me the right direction. There is a lot to read. Let me finish it. 

 

Edit. I suspect word too. A normal blank template can show all my letters without any problem. But a specific template which i work on is causing this problem.

Edited August 9, 2015 by kcvinu

[jchd]

Is the offending template explicitely tagged as "malayalam"? If it's English (say), things may not work as expected, but this is just a guess.

This site (by Richard Ishida, a lead member of the Unicode consortium and I8n) is full of up-to-date data and information about Unicode. For instance its uniview applet lets you see a Unicode block (e.g. malayalam) and access detailed information about each and every codepoint (e.g. 0x0D4D).

EDIT: I must clarify something for future readers. When I or Ishida use the word "script", we intend "a writing system" which can be Latin, Cyrillic, Hangul, ... and this has no relationship with the term "script" as in "AutoIt is a script language".

Edited August 9, 2015 by jchd