SnipIt Options

[ghostwalker]

//edit: here's another example using a 3rd party pasteboard manager…

That wood make a cool Widget also.

As for the crop tool, still working on it, but because the window is limited in size, I have to make it expandable after it is set, so you guys are getting something similar to the later GIMPs' selection tool. Lucky you guys!

That is very cool cant wait to check it out.

[dimsum]

… making a Widget clone of a really neat program (a damn good clone, if I do say so myself)

It is a damn fine clone!

… intended to be a quick "grab the image, short edit, save"

Didn't mean to throw all those feature requests at you, nor was I trying to hold you to stand-alone app standards. You wrote a really terrific widget!

(WTF with these boards - everytime I try to submit a reply I'm logged out!)

[Hunter890]

Hey, I tried out my idea and it wasn't half bad!

[ghostwalker]

That's really nice I like it.

[CSM]

Edit: nevermind.

This post has been edited by CSM: Jan 29 2009, 02:51 AM

[CSM]

Version 1.1.0 is up.

Change Log:
- Made title-bar text more robust (won't throw errors if clicked), font now set to tahoma, size 11, and bold
- Fixed edit window button spacing (1px)
- Added text to undo button
- Changed undo button and highlight tool's icons to better suit
- Combined the current pencil color with the "change color"
- Changed over drawing stuff to use a current variable, instead of top value on the undo-stack.
- All drawing methods now change values on already existing Point objects, instead of instanciating an unnecessary object each mouse move.
(Free-hand/form, and the Pencil tool require a point each move, so it can't be avoided in those cases)
- Rectangular screenshot capture now listens for shift and snaps to a square selection while dragging
- Added space for help text about tools (used by crop and annotation)

- Implemented cropping tool with flexible selection
- Added annotation tool, for text


Some notes:

Clicking the menu item for various colors or fonts will bring up the color dialog (as before with the "change pencil color" option).

The crop tool works like this: click the crop button, select an area, resize selection area as needed, then hit the crop button again. To cancel a crop-selection or undo an entire crop, just hit the undo button or any other tool's button. The cropping tool does support "square" selections, just hold down the shift key while dragging, and it will snap the selection's aspect to a square based on the x-axis (for initial placement), or whichever you are changing (resizing).

The annotation tool works sorta like the crop tool. You select a rectangular area, which is then filled with a TextArea when you let up on the mouse. When you hit the annotation button again it will close the text and add it to the image. The undo button will remove any annotations or texts in progress. Do note that the background (if enabled) is only as large as the selected area, so you need to make it bigger than your text. You can change the text color, background color, and font used. If you cause it to have a scrollbar, the scrollbar is not captured in the output, so scroll what you want visible before making it permanent.


I've reached the amount of features I would prefer to stay at. Given the state of the code (it started small, and grew big in a hurry, and I never reorganized it) it's getting harder to me to find the little things (even though I wrote it!) and as such, I'm more likely to forget something or break something else. Who knew that something which looks so good has messy code? Some of this would have been easier/shorter if I didn't have to create all my own buttons, but I don't think that if they were built in, I'd get the look/control I want anyways ... but that's how it goes: you get good enough you end up designing your own stuff (or extending built-ins) because nothing is "perfect"* (long live Swing? ).

* I'm sure most good GUI designers know what I mean, you're always after that extra little bit, like turning off the focus-border on JTableCells.

[g6auc]

The Text tool is not working for me. I can select the tool, mark out a rectangle and type the text. When i click the tool again, the text disappears. I am also seeing this:

CODE

Loaded Widget 'SnipIt' from /Users/xxx/Desktop/SnipIt.widget
runCommandInBg: java -jar "/Users/xxx/Library/Application Support/Yahoo! Widget Engine/Widgets/SnipIt.widget/clipboard.jar" "/Users/xxx/Library/Application Support/Yahoo! Widget Engine/Widgets/SnipIt.widget/edit.png" true tag: copy
runCommandInBg: java -jar "/Users/xxx/Library/Application Support/Yahoo! Widget Engine/Widgets/SnipIt.widget/clipboard.jar" "/Users/xxx/Library/Application Support/Yahoo! Widget Engine/Widgets/SnipIt.widget/edit.png" true tag: copy
TypeError: current has no properties (main.js: Line 1209)

The help field on the bottom of the control window is being overwritten.

[CSM]

I am honestly not seeing that error ("current has no ..."). I even deleted the reg key and Widget data folder stuff to start.

Version 1.1.1 is up.

Change Log:
- Default annotation text is now black.

[g6auc]

I am honestly not seeing that error ("current has no ..."). I even deleted the reg key and Widget data folder stuff to start. Version 1.1.1 is up. Change Log:- Default annotation text is now black.

I'm seeing the black text, and the same problems - the text disappears when I click text a second time after typing. I get the same error - click undo a few times, then click text.



Also seeing this:


If it's any help - there are a number of images (called textNN.png) in the data folder. They are of reasonable size, but completely blank and transparent.



This post has been edited by g6auc: Feb 3 2009, 01:00 AM

[dimsum]

I am honestly not seeing that error ("current has no ..."). I even deleted the reg key and Widget data folder stuff to start.

Version 1.1.1 is up.

Change Log:
- Default annotation text is now black.

Same as g6auc on my Mac. Additionally, sometimes when clicking on the text box, I get an image-loading error:

CODE

runCommandInBg: java -jar "/Users/***/Library/Application Support/Yahoo! Widget Engine/Widgets/SnipIt1.1.1.widget/clipboard.jar" "/Users/***/Library/Application Support/Yahoo! Widget Engine/Widgets/SnipIt1.1.1.widget/edit.png" tag: copy
TypeError: current has no properties (main.js: Line 1209)
KFImageView: Can't load image "/Users/***/Library/Application Support/Yahoo! Widget Engine/Widgets/SnipIt1.1.1.widget/text0.png"

[ghostwalker]

Great update, I have not played with it yet but I will after work.

[CSM]

Version 1.1.2 up.

Change Log:
- Fixed issue where button help text was not cleared when edit window was closed.
- Fixed issue where not having a tool in the current variable caused errors.
- Added a sleep(1) after text is saved out to the file.

[dimsum]

Version 1.1.2 up.

Change Log:
- Fixed issue where button help text was not cleared when edit window was closed.
- Fixed issue where not having a tool in the current variable caused errors.
- Added a sleep(1) after text is saved out to the file.

CODE

KFImageView: Can't load image "/Users/***/Library/Application Support/Yahoo! Widget Engine/Widgets/SnipIt1.1.2.widget/text4.png"

[CSM]

Yes, that error was reported previously by Hunter890. There's not a whole lot I can do about this. Widgets cannot have two events fire at the same time, so it should follow that code after a saveImageToFile call cannot be executed until the image is fully saved. In practice, however, this is not what happens, and the saveImageToFile call returns just before the image is fully written to disk on slower systems. This leaves me to believe that despite attempts to handle all events on a single thread behind the scenes, some things are done asynchronously, which messes with stuff depending on it to be blocking.

Anyways, "not a whole lot" means I can make an attempt, but not a good one.

Version 1.1.3 is up.

Change Log:
- Increased the sleep count after the text is saved to an extremely long amount of 250 ms.

[dimsum]

Yes, that error was reported previously by Hunter890. There's not a whole lot I can do about this. Widgets cannot have two events fire at the same time, so it should follow that code after a saveImageToFile call cannot be executed until the image is fully saved. In practice, however, this is not what happens, and the saveImageToFile call returns just before the image is fully written to disk on slower systems. This leaves me to believe that despite attempts to handle all events on a single thread behind the scenes, some things are done asynchronously, which messes with stuff depending on it to be blocking.

Anyways, "not a whole lot" means I can make an attempt, but not a good one.

Version 1.1.3 is up.

Change Log:
- Increased the sleep count after the text is saved to an extremely long amount of 250 ms.

I'm no longer getting errors in the debug window, but the text still won't save or get copied to clipboard with the original capture the way it does with the highlight or pencil tools. Thanks for giving it a go, though. (Perfectly understandable if you want to drop this feature.)

[CSM]

I'm no longer getting errors in the debug window, but the text still won't save or get copied to clipboard with the original capture the way it does with the highlight or pencil tools. Thanks for giving it a go, though. (Perfectly understandable if you want to drop this feature.)

The other tools work because I can draw them directly to the Canvas. The text tool must be save to an image first because there is no saveImageToImage(*) to return an image within Y!WE ... I have to use the fileystem.

[dimsum]

The other tools work because I can draw them directly to the Canvas. The text tool must be save to an image first because there is no saveImageToImage(*) to return an image within Y!WE ... I have to use the fileystem.

OHH! Now I understand

[g6auc]

I'm still getting problems with 1.1.3.

When I click the Text button to save a text, a file appears immediately in the data folder but it is blank and transparent.
So it looks to me as if the creation of the image of the text is at fault, and (probably) not the use of that image of the text.

I've used the same trick myself to write text to a canvas, so I know that it can be done.

EDIT: In my CanvasTextCode, I use txt.saveImageToFile( path, "png" ); where txt is a (single line) Text object.

I can't remember if I tried it with TextArea objects. You could try copying the text from the textarea to a multi-line text object with wrap set, to see if that works.

I'll try to make a little test program later today, to see if I can save TextAreas on the Macintosh.

EDIT: I've run a little test in WidgetTester which confirms that textarea.saveImageToFile( path, "png" ); does not work correctly on the Macintosh. An image is written, but it is blank and totally transparent.

CODE

var saveTextArea = function () {
    var path = "/Users/xxx/Desktop/textarea.png";
    print(command.data);
    command.saveImageToFile( path, "png" );
};

then call it form the command line with saveTextArea();.

EDIT: OK - this works:

CODE

var saveTextArea = function (textarea) {
    var path = "/Users/nhw/Desktop/textarea.png";
    var txt = new Text();
    txt.wrap = true;
    txt.data = textarea.data;
    txt.saveImageToFile( path, "png" );
};

This produces an image containing the text. It would be wise to define more of the txt attributes to set the style, background, etc. These could be copied from the attributes of the textarea object.

On the Macintosh, I'm not seeing backgrounds on the screen when I create a Text with the Text button. Do you set the color, opacity etc?




This post has been edited by g6auc: Feb 3 2009, 10:40 AM

[CSM]

To answer all present and future* responses and questions:
* about the fixes

- I forgot about that bug! ****it, this is a perfect example of why programs like this shouldn't be written with two different codebases (we have cross-platform languages/GUI-toolkits now, you know!)
- Yes I set the opacity, but I used the wrong preference, so it actually set it to a hex color, not a number.
- TextAreas actually inset their text by 3 pixels from the top, and 6 pixels from the edges, so I adjusted for that to ensure text doesn't wrap different and doesn't move.
- Text objects don't have as crisp text as TextAreas do.


Version 1.1.4 is up.

Change Log:
- Annotation tool now uses a Text object to capture the text to be added to the edit image.
- Default background for the annotation tool is now an off-white to match my change of the default foreground color to black.
- Decreased the sleep time between capturing the image and using it to 100 ms for the annotation tool (hope it still works).



I do have a question: does the non-mention of the crop tool mean that it works?

[ghostwalker]

Works fine here on Windows.