SRL Archive - RS has been updated.

Hello,

I heard there are a few new uptext colours. I'd like everyone to post a few pictures (only of the uptext, not the entire screen) of the uptext so we can tweak our current algorithm.

You can upload them to http://imgur.com

Please keep them as bmp or png, no jpg.

If we have quite some, perhaps someone can throw ACA at each of the different colours so we can get a nice range for each colour.

EDIT:

So right now I've set up autobuild for OCRBench. Temporary link: http://nala.villavu.com/downloads/master/5b5511fff5b7fd3c4388cfc4d655ca2e6c5b48fd/
Eventually I think OCRBench should allow the user to customise the colour filter so that someone other than me can experiment and fix the uptext.

If anyone feels like working on OCRBench (as well as changing the OCR algorithms), be my guest. I probably won't fix this tonight.

EDIT2:

Have a look at Units/MMLCore/ocr.pas line 240..354. That's pretty much the colour filter. All that needs to be done is adding/changing the colour filter there.

Do it yourself noob.
:)

From: YoHoJo
Do it yourself noob.
:)

Will you fix the uptext yourself?

Sorry if it's too small. Do you need anything particular?

Get all the different colors you can find. There's a new grey i know, and a gold/grey etc.

E:
Image loaded slow… that the only new colors so far?

From: kingarabian

Sorry if it's too small. Do you need anything particular?

I would have very much appreciated if it was not all on a single bitmap, but I think the size is alright. I had forgotten the text was so thin. :p

http://prntscr.com/5rg2l
http://prntscr.com/5rg2w
http://prntscr.com/5rg34
http://prntscr.com/5rg3b

From: Narcle
Get all the different colors you can find. There's a new grey i know, and a gold/grey etc.

E:
Image loaded slow… that the only new colors so far?

The green uptext that shows when you mouse an item in your inventory looks new. Or is it gray? I dunno.

Magic spells uptext:

Question.. with all those pictures, the font is the same, the colours change.. is it not possible to have simba load fonts with transparent backgrounds and black writing then have it just match the shape instead of the letters? Because everytime a new colour comes out that doesn't fit our tolerances, fonts will break..

Problem is the background can be same/near same color as UpText, and its never the same color.

From: ggzz
Question.. with all those pictures, the font is the same, the colours change.. is it not possible to have simba load fonts with transparent backgrounds and black writing then have it just match the shape instead of the letters? Because everytime a new colour comes out that doesn't fit our tolerances, fonts will break..

If you can find a way to detect shapes that are that tiny in a reasonably fast manner, let me know. It's probably possible - but it takes time, and that's unfortunately one thing I don't have a lot of right now.

Here are some more colored uptexts

http://i.imgur.com/YV8jc.png

and
http://i.imgur.com/TyFII.png

http://imgur.com/9kZND,IibXn,HL8Tf,N59Nx,kPPEv

EDIT: Make sure to check all five images (at the top).

From: Wizzup?
If you can find a way to detect shapes that are that tiny in a reasonably fast manner, let me know. It's probably possible - but it takes time, and that's unfortunately one thing I don't have a lot of right now.

I'll try.. my idea is to actually write a font finder really.. basically I'm going to write a script that will save tpa's of the top left box of the screen.. Rotate, then save.. whatever pixels stay exactly the same, is the uptext.. repeat as many times as needed for accuracry.. set those pixels white and the background black..

Updated first post. I'm not going to fix it tonight, because I have to get up really early tomorrow. I can probably have it fixed in about 24 hours from now, but it would really nice if someone else can either completely fix it (it's not too hard if you can compile OCRBench/Simba with lazarus) or at least compile a list of colours. (Several colours per ``type'' of colour)

Cheers.

E: http://l0.lt now also builds OCRBench.

I know it's probably a long shot, but have we ever considered using a font of the shadow to find uptext? I know there are gaps, but surely that can be resolved.

program new;
{$i srl/srl.simba}

function getUptextTPA(var tpa: TPointArray): boolean;
begin
  result := findColorsTolerance(tpa, 132357, 5, 5, 260, 30, 50);
end;

procedure debugUptext();
var
  i, w, h, bmp: integer;
  tpa: TPointArray;
begin
  w := (MSX1 + 260);
  h := (MSY1 + 30);

  bmp := createBitmap(w, h);

  fastDrawClear(bmp, clBlack);

  if (getUptextTPA(tpa)) then
  begin
    for i := 0 to high(tpa) do
      fastSetPixel(bmp, tpa[i].x, tpa[i].y, clRed);

    displayDebugImgWindow(w, h);
    drawBitmapDebugImg(bmp);
  end;
end;

var
  tpa: TPointArray;
begin
  clearDebug();
  activateClient();
  setupSRL();

  mmouse(664, 226, 5, 5);
  wait(500);

  debugUptext();
end.

Afaik uptext finding already uses shadow.

From: Markus
Afaik uptext finding already uses shadow.

I know, but it uses colors as well. I'm wondering if it's possible to look just for the shadow text, and ignore the colored part.

Here we go…


That's another color that you guys might be interested in.

From: Coh3n
I know, but it uses colors as well. I'm wondering if it's possible to look just for the shadow text, and ignore the colored part.

Think this is an absolutely brilliant idea. I bet y'all could make this work. It seems fairly logical and almost fail-proof if coded correctly.

Seriously, genius idea. :cool:

If someone knows a way to clear up the pictures created, lemmie know..

I used the below code.. gets colours in the box 4 times.. checks which pixels changed and turns those ones black.. The text pixel colours are wayy brighter and have a way lower tolerance than the background when rotated.. if u can find pixels of a specific brightness tolerance then that'll improve the accuracy of this.. I thought it'd be easy to dynamically make fonts but it isn't..

  {$DEFINE SRL5}
  {$DEFINE CRASHSMART}
  {$DEFINE SMART}
  {$I SRL/srl.Simba}
  {$I SRL/SRL/Misc/Debug.Simba}

type
  TPAC = record
    TPA: TPointArray;
    TIA: TIntegerArray;
  end;
(*
Function GetTPACBox(B: TBox): TPAC;
var
  Temp: TPAC;
  TP: TPoint;
  I, J, K, W, H: Integer;
begin
  W:= B.X2 - B.X1;
  H:= B.Y2 - B.Y1;

  K:= 0;
  For I:= 0 To H do
  begin
    For J:= 0 To W do
    begin
      SetLength(Temp.TPA, Length(Temp.TPA) + 1);
      TP:= Point(B.X1 + J, B.Y1 + I);
      Temp.TPA[K]:= TP;
      Inc(K);
    end;
    J:= 0;
  end;
  SetLength(Temp.TIA, Length(Temp.TPA));
  Temp.TIA:= GetColors(Temp.TPA);
  Result:= Temp;
end;
*)

Function GetTPACBox(B: TBox): TPAC;
var
  Temp: TPac;
begin
  Temp.TPA:= TPAFromBox(B);
  Temp.TIA:= GetColors(Temp.TPA);
  Result:= Temp;
end;

Procedure GetTextBoxes;
var
  Temp: Array of TPAC;
  B: TBox;
  I, K, BMP, W, H: Integer;
begin
  B:= IntToBox(7, 10, 285, 27);
  SetLength(Temp, 4);
  Temp[0]:= GetTPACBox(B);
  Sleep(1000);
  MakeCompass('E');
  Sleep(1000);
  Temp[1]:= GetTPACBox(B);
  MakeCompass('S');
  Sleep(1000);
  Temp[2]:= GetTPACBox(B);
  MakeCompass('W');
  Sleep(1000);
  Temp[3]:= GetTPACBox(B);

  W:= B.X2 - B.X1;
  H:= B.Y2 - B.Y1;
  K:= 0;
    For I:= 0 To High(Temp[0].TIA) do
    begin
      if(
      Not(SimilarColors(Temp[0].TIA[K], Temp[1].TIA[K], 25)) or
      Not(SimilarColors(Temp[0].TIA[K], Temp[2].TIA[K], 25)) or
      Not(SimilarColors(Temp[0].TIA[K], Temp[3].TIA[K], 25)) or
      Not(SimilarColors(Temp[1].TIA[K], Temp[2].TIA[K], 25)) or
      Not(SimilarColors(Temp[1].TIA[K], Temp[3].TIA[K], 25)) or
      Not(SimilarColors(Temp[2].TIA[K], Temp[3].TIA[K], 25))
      ) then
      begin
        Temp[0].TIA[I]:= 0;
        Temp[1].TIA[I]:= 0;
        Temp[2].TIA[I]:= 0;
        Temp[3].TIA[I]:= 0;
      end;
      Inc(K);
    end;

  BMP:= CreateBitmap(W + 50, H + 50);

  FastSetPixels(BMP, Temp[0].TPA, Temp[0].TIA);

  DebugBitmap(BMP);
  FreeBitmap(BMP);
end;

begin
  Smart_Server := 72;
  Smart_Members := True;
  Smart_Signed := True;
  Smart_SuperDetail := False;
  setupSRL;
  MMouse(664, 228, 0, 0);
  GetTextBoxes;
end.

From: Coh3n
I know, but it uses colors as well. I'm wondering if it's possible to look just for the shadow text, and ignore the colored part.

This is possible, but I don't know why you'd want to do that. If anything, searching for the colours as well highly increases the accuracy of the algorithm.
See: http://docs.villavu.com/simba/mmlref/ocr.html#uptext

I'll work on this tonight.

From: Wizzup?
This is possible, but I don't know why you'd want to do that. If anything, searching for the colours as well highly increases the accuracy of the algorithm.
See: http://docs.villavu.com/simba/mmlref/ocr.html#uptext

I'll work on this tonight.

I'd want to do it so we'd never have this issue again, unless Jagex changed the shadow color to, say, red.

OK. I think I added the gray colour succesfully. I haven't tested it extensively beyond on the images on this thread. If it fails on a particular uptext, take a picture and post it here.

Find Simba 983 here: http://nala.villavu.com/downloads/simba-prev/61edd4a0767131980e3a1f5756a5f9877d6df799/Simba.i386-win32.exe

Please test and tell me if it works. I am afraid that walls might become a more serious ``noise'' problem since I added a somewhat wide range of gray.

I will test it more extensively later. For now, go forth and test! And of course, macro. :spot:

E: All downloads of that revision here: http://nala.villavu.com/downloads/simba-prev/61edd4a0767131980e3a1f5756a5f9877d6df799/

HELL YEAH!
Thank you Wizzup, all helpers, and pic posters!

E:
Download links messed up?
I just get a bunch of wierdly formatted font stuff of ????s when I click download links.


E:
I used Jenkins and got Simba (Next) and it's Rev 1000.
This will work, or nah?

From: YoHoJo
HELL YEAH!
Thank you Wizzup, all helpers, and pic posters!

E:
Download links messed up?
I just get a bunch of wierdly formatted font stuff of ????s when I click download links.


E:
I used Jenkins and got Simba (Next) and it's Rev 1000.
This will work, or nah?

I no longer recommend jenkins atm, since I've been working on integrating buildbot to take off some pressure of MITB and allow myself to build more branches as tools as I see fit. But it should be in Simba (prev) on Jenkins as well.

Not finding.

From: kingarabian

Not finding.

Works perfectly for me in the OCRBench? (And in Simba too, then - same code)

Not sure but it not reading this:
Colours I got:
7456741
7129320
7063268
6010322
7193826
Mat

From: kingarabian

Not finding.

doesn't find:
<– only found if uptext search is ark.
<– only found if uptext search is yip pou
<– only found if uptext search is moning po

Guys, try:

writeln(rs_getuptext);

From: Mat
Not sure but it not reading this:
Colours I got:
7456741
7129320
7063268
6010322
7193826
Mat

It finds everything except the 'sh' in ashes for me. (I know what's causing it and I know how to fix it - probably.) But it'll require a partial rewrite of the algorithm. I won't do that tonight. It should at least pick up the gray colours now though.

Ow :P and I am using ashes. That explains it :)
Mat

HLR works like charm once again :smile: <3

From: ggzz
If someone knows a way to clear up the pictures created, lemmie know..

I used the below code.. gets colours in the box 4 times.. checks which pixels changed and turns those ones black.. The text pixel colours are wayy brighter and have a way lower tolerance than the background when rotated.. if u can find pixels of a specific brightness tolerance then that'll improve the accuracy of this.. I thought it'd be easy to dynamically make fonts but it isn't..

  {$DEFINE SRL5}
  {$DEFINE CRASHSMART}
  {$DEFINE SMART}
  {$I SRL/srl.Simba}
  {$I SRL/SRL/Misc/Debug.Simba}

type
  TPAC = record
    TPA: TPointArray;
    TIA: TIntegerArray;
  end;
(*
Function GetTPACBox(B: TBox): TPAC;
var
  Temp: TPAC;
  TP: TPoint;
  I, J, K, W, H: Integer;
begin
  W:= B.X2 - B.X1;
  H:= B.Y2 - B.Y1;

  K:= 0;
  For I:= 0 To H do
  begin
    For J:= 0 To W do
    begin
      SetLength(Temp.TPA, Length(Temp.TPA) + 1);
      TP:= Point(B.X1 + J, B.Y1 + I);
      Temp.TPA[K]:= TP;
      Inc(K);
    end;
    J:= 0;
  end;
  SetLength(Temp.TIA, Length(Temp.TPA));
  Temp.TIA:= GetColors(Temp.TPA);
  Result:= Temp;
end;
*)

Function GetTPACBox(B: TBox): TPAC;
var
  Temp: TPac;
begin
  Temp.TPA:= TPAFromBox(B);
  Temp.TIA:= GetColors(Temp.TPA);
  Result:= Temp;
end;

Procedure GetTextBoxes;
var
  Temp: Array of TPAC;
  B: TBox;
  I, K, BMP, W, H: Integer;
begin
  B:= IntToBox(7, 10, 285, 27);
  SetLength(Temp, 4);
  Temp[0]:= GetTPACBox(B);
  Sleep(1000);
  MakeCompass('E');
  Sleep(1000);
  Temp[1]:= GetTPACBox(B);
  MakeCompass('S');
  Sleep(1000);
  Temp[2]:= GetTPACBox(B);
  MakeCompass('W');
  Sleep(1000);
  Temp[3]:= GetTPACBox(B);

  W:= B.X2 - B.X1;
  H:= B.Y2 - B.Y1;
  K:= 0;
    For I:= 0 To High(Temp[0].TIA) do
    begin
      if(
      Not(SimilarColors(Temp[0].TIA[K], Temp[1].TIA[K], 25)) or
      Not(SimilarColors(Temp[0].TIA[K], Temp[2].TIA[K], 25)) or
      Not(SimilarColors(Temp[0].TIA[K], Temp[3].TIA[K], 25)) or
      Not(SimilarColors(Temp[1].TIA[K], Temp[2].TIA[K], 25)) or
      Not(SimilarColors(Temp[1].TIA[K], Temp[3].TIA[K], 25)) or
      Not(SimilarColors(Temp[2].TIA[K], Temp[3].TIA[K], 25))
      ) then
      begin
        Temp[0].TIA[I]:= 0;
        Temp[1].TIA[I]:= 0;
        Temp[2].TIA[I]:= 0;
        Temp[3].TIA[I]:= 0;
      end;
      Inc(K);
    end;

  BMP:= CreateBitmap(W + 50, H + 50);

  FastSetPixels(BMP, Temp[0].TPA, Temp[0].TIA);

  DebugBitmap(BMP);
  FreeBitmap(BMP);
end;

begin
  Smart_Server := 72;
  Smart_Members := True;
  Smart_Signed := True;
  Smart_SuperDetail := False;
  setupSRL;
  MMouse(664, 228, 0, 0);
  GetTextBoxes;
end.

That doesn't look too bad, but doesn't this require changing the compass for example. Or are you only doing this to find the colours? Could you try to explain the algorithm in a few sentences?

From: Wizzup?
That doesn't look too bad, but doesn't this require changing the compass for example. Or are you only doing this to find the colours? Could you try to explain the algorithm in a few sentences?

i can
it makes a array of the position and colour of each pixel, moves the compass and does it again

then after it compares the pixels and blacks out all the changed ones

however this method requires the background to move

~shut

From: Shuttleu
i can
it makes a array of the position and colour of each pixel, moves the compass and does it again

then after it compares the pixels and blacks out all the changed ones

however this method requires the background to move

~shut

Yeah.. Basically I logged in my character, ran the script.. It got a TPA of the uptext box. Rotated the camera and did that again. Does it 4 times. Since the uptext stays the same, and the background changes colours all the time, it turns the changed pixels to black. Leaving the uptext there :) Only works sometimes though.. in places like the desert where the background is always the same, it won't work.

From: Coh3n
I'd want to do it so we'd never have this issue again, unless Jagex changed the shadow color to, say, red.

How will it work in dungeons(or places with black background)? Because the colors will all look the same and then only thing you have is the color of the uptext.

From: ggzz
Yeah.. Basically I logged in my character, ran the script.. It got a TPA of the uptext box. Rotated the camera and did that again. Does it 4 times. Since the uptext stays the same, and the background changes colours all the time, it turns the changed pixels to black. Leaving the uptext there :) Only works sometimes though.. in places like the desert where the background is always the same, it won't work.

What if the object is tiny and changing the compass changes the uptext?

From: KingKong
How will it work in dungeons(or places with black background)? Because the colors will all look the same and then only thing you have is the color of the uptext.




What if the object is tiny and changing the compass changes the uptext?

It's used to grab fonts.. and to make it easier to create.. Not for grabbing actual uptext in a script.. Sucks I guess lol Wiz is working on other methods so we'll have to see how that goes.

From: ggzz
It's used to grab fonts.. and to make it easier to create.. Not for grabbing actual uptext in a script.. Sucks I guess lol Wiz is working on other methods so we'll have to see how that goes.

O.o I thought you meant it was used to grab the uptext. Im pretty sure you can get a bmp of all the fonts using opengl or something and slice them up.

From: KingKong
How will it work in dungeons(or places with black background)? Because the colors will all look the same and then only thing you have is the color of the uptext.

In that case you could return every point that isn't black.

From: Coh3n
In that case you could return every point that isn't black.

But its not always possible to know if we are inside the dungeon. Also what if half the uptext is inside the black area, and the other half is in 'normal' area(as in the ground or something).

Anyway here's my idea:
1)Get all the pixels in the uptext area
[ATTACH]13599[/ATTACH]

2)Convert the colors to monochrome
[ATTACH]13600[/ATTACH]

3)Search for font(or rather partial search)

From: KingKong
But its not always possible to know if we are inside the dungeon. Also what if half the uptext is inside the black area, and the other half is in 'normal' area(as in the ground or something).

You're right. :p Would just be nice to be able to get the uptext without colors so when Jagex changes colors, it doesn't effect us. :)

From: Coh3n
You're right. :p Would just be nice to be able to get the uptext without colors so when Jagex changes colors, it doesn't effect us. :)

Thats what the rest of my post was about, it doesn't require getting colors or anything and if done correctly, it should actually be faster than the one we're using right now, but im not sure how well it would work with darker colors(since they might be turned black with the background).

so is the bot working or is that wat we are trying to fix? im new :D i wanna learn to write my own scripts aswell
:)

coz when i open the MSI it goes to a paintsmart insted of opening the bot character settings thingyy lol

It struggles beyond belief with the grey uptext in Frost Dragon Lair.

It seems to believe it can find ANYTHING. I will try and post a screenie tonight but I am supposedly working and fell asleep at my computer last night when attempting to.

Edit: Sorry my idea of a fix would be a decrease in the Tolerance. Though I know this would cause more problems than most things.

Meh. I open it and it says i have no extensions.

From: KingKong
Thats what the rest of my post was about, it doesn't require getting colors or anything and if done correctly, it should actually be faster than the one we're using right now, but im not sure how well it would work with darker colors(since they might be turned black with the background).

Results:

On white backgrounds, the text shows clearly to my eyes but for simba it may be a problem.. Not much of a problem though because if you combine it with the background rotation and set the changed pixels to black, it'll be so epic.. Yeah.. I've tested it with the background rotation and it works.. This could only be used to MAKE fonts though.. not to FIND them as it will fail miserably in icy places.

Does exactly what you suggested (no rotation, easy to implement):

  {$DEFINE SRL5}
  {$I SRL/srl.Simba}
  {$I SRL/SRL/Misc/Debug.Simba}


Procedure GetBoxes(var TPA: TPointArray; var TIA: TIntegerArray; B: TBox);
begin
  TPA:= TPAFromBox(B);
  TIA:= GetColors(TPA);
end;

Procedure CreateBitmapFont(TPA: TPointArray; TIA: TIntegerArray; B: TBox);
var
  BMP, D, BN, C, W, H: Integer;
begin
  W:= B.X2 - B.X1;
  H:= B.Y2 - B.Y1;
  BMP:= CreateBitmap(W + 25, H + 25);

  FastSetPixels(BMP, TPA, TIA);
  //D:= DesaturateBitmap(BMP);   //15000804 = white writing..
  D:= GreyScaleBitmap(BMP);      //Desat is darker than greyscale..
  BN:= BrightnessBitmap(D, 2);
  C:= ContrastBitmap(BN, 5);


  DebugBitmap(c);

  FreeBitmap(BN);
  FreeBitmap(C);
  FreeBitmap(D);
  FreeBitmap(BMP);
end;

var
  B: TBox;
  TPA: TPointArray;
  TIA: TIntegerArray;


begin
  setupSRL
  B:= IntToBox(6, 9, 276, 26);
  GetBoxes(TPA, TIA, B);
  CreateBitmapFont(TPA, TIA, B);
end.

From: Mat
Not sure but it not reading this:
Colours I got:
7456741
7129320
7063268
6010322
7193826
Mat

This is fixed in http://villavu.com/forum/showthread.php?t=73988

E: ggzz remind me to look at your function in more detail. Let me just tell you that you're the first I know of to actually use those greyscale and brightness functions. I actually tried that for previous OCR attempts btw. It didn't prove to be accurate enough, but if you can make it accurate, I'm very interested. Right now the uptext plus ocr only takes like ~16ms, so we've got plenty of room for some more accuracy.

From: ggzz
This could only be used to MAKE fonts though.. not to FIND them as it will fail miserably in icy places.

I just tested at the dungeoneering place where the character is teleported to, and here is what i get:

[ATTACH]13621[/ATTACH]

And im pretty sure its not that hard to make it find them as well.

From: KingKong
I just tested at the dungeoneering place where the character is teleported to, and here is what i get:

[ATTACH]13621[/ATTACH]

And im pretty sure its not that hard to make it find them as well.

Yeah it'll do that on the snow places because the background of that dung place is pure white as is the text when you contrast it.. I tried it at castlewars and mobilizing armies and a couple other places and it worked pretty nicely. But Ice places are definitely a no-no. Actually gunna work on it again and see if there is another way to exclude the backgrounds or somehow detect what is what.

From: Coh3n
You're right. :p Would just be nice to be able to get the uptext without colors so when Jagex changes colors, it doesn't effect us. :)

Hmm, my algorithm in SCAR has worked since like 2009 and still does, only change I've had to make it allowing for multiple fonts to be used, but nothing to actually change the algorithm itself…

From: Freddy1990
Hmm, my algorithm in SCAR has worked since like 2009 and still does, only change I've had to make it allowing for multiple fonts to be used, but nothing to actually change the algorithm itself…

What is it exactly you are proffering here? That you might help out with the algorithm used by Simba, or just letting us know that yours works? :unsure:

From: Freddy1990
Hmm, my algorithm in SCAR has worked since like 2009 and still does, only change I've had to make it allowing for multiple fonts to be used, but nothing to actually change the algorithm itself…

Dunno about that. I used SCAR last night and it wouldn't detect anything.

From: Freddy1990
Hmm, my algorithm in SCAR has worked since like 2009 and still does, only change I've had to make it allowing for multiple fonts to be used, but nothing to actually change the algorithm itself…

That's great, but unless you released the source code it isn't much help to us. :o

From: Coh3n
That's great, but unless you released the source code it isn't much help to us. :o

Hence why he came on here to show off :shifty:

From: Freddy1990
Hmm, my algorithm in SCAR has worked since like 2009 and still does, only change I've had to make it allowing for multiple fonts to be used, but nothing to actually change the algorithm itself…

Actually, when Simba wasn't even released I made a plugin with the Simba OCR that would load in SCAR and make my scripts work again, because the SCAR algorithm was broken for several weeks straight… I really wanted to release it at the time but decided it wouldn't be a nice/smart move… :p

Anyway, I'm looking forward to a pull request on the Simba repository with your ``better'' algorithm. :) Unless you of course really don't intend to share how it works with anyone - in which case the entire post is redundant and quite sad - you make me sad with this kind of posts.

E: I think I kind of remember how it worked, btw. Using shadow is still using colour, which is also what Simba uses.

Freddy please see the light- Open Source is the future. I mean even the worlds fastest and strongest super computer uses Linux.

Back on topic, hopefully we won't have to worry about stupid uptext updates breaking our scripts with this Simba update…

From: kingarabian
Freddy please see the light- Open Source is the future. I mean even the worlds fastest and strongest super computer uses Linux.

Back on topic, hopefully we won't have to worry about stupid uptext updates breaking our scripts with this Simba update…

If the uptext changes - you'll generally need an update somewhere. We can't write code for something we have never seen before. ;)
It really all depends on how much they change the (up)text. I can see a lot of future roadblocks in this area.

From: Wizzup?
If the uptext changes - you'll generally need an update somewhere. We can't write code for something we have never seen before. ;)
It really all depends on how much they change the (up)text. I can see a lot of future roadblocks in this area.

What if just the color changes, not the font?

From: kingarabian
What if just the color changes, not the font?

Then you can set it in SRL (in the near future). See http://villavu.com/forum/showthread.php?p=913586#post913586
So that'll only require a SRL update to the filters.

can someone please help me get mine working? answer me atleast :(?

Not sure what the links are about/why I can't seem to download newest version, but will check back soon.

I was thinking of an overlay type thing?
Basically in the area where the text is defined if it cant recognise the text a backup script should run and overlay possible words to see if it can find a match. E.g. Have a fluro pink abc kind of algorith run over the letters till it covers and matches all of the letters.
Not as a main descrepancy but as a backup.

Say for eg

Summoning pouch
The normal method picks up "moning pou"
That could be run through a list of words untill it sits perfectly over the word
Summoning pouch hiding the fluro pink and detecting a match.

For example if the script was looking for Summoning Pouch
What it would do is Screen cap the small area around where the words should be and put it over a pink version of the same word.
If all the pink is hidden match.
If not It keeps looking?

By the way i was wondering if i could get a few links to Some SRL 5.0 guides
As im quite interested and i think i would make a nice script for "ardy" (I think its ardy?" mine to add to msi.
I know you're gunna be all like >:google:
But i would appreciate the help.

MSI: http://villavu.com/forum/showthread.php?t=56614