I'm using a TreeView in GTK# and now I need to use a custom cell renderer.
Although I got almost everything to work, there's one missing piece: The cell renderer must know what to render.
The data to render is a custom class. I want to retrieve a field from that class.
Here's what I have so far
public class TreeItemCellRenderer : CellRenderer
{
public override void GetSize (Widget widget, ref Gdk.Rectangle cell_area, out int x_offset, out int y_offset, out int width, out int height)
{
base.GetSize (widget, ref cell_area, out x_offset, out y_offset, out width, out height);
height = 16;
}
protected override void Render (Gdk.Drawable window,
Widget widget, Gdk.Rectangle background_area, Gdk.Rectangle cell_area, Gdk.Rectangle expose_area, CellRendererState flags)
{
base.Render (window, widget, background_area, cell_area, expose_area, flags);
using (var layout = new Pango.Layout(widget.PangoContext)) {
layout.Alignment = Pango.Alignment.Left;
layout.SetText("Hello, World!");
StateType state = flags.HasFlag(CellRendererState.Selected) ?
widget.IsFocus ? StateType.Selected : StateType.Active : StateType.Normal;
window.DrawLayout(widget.Style.TextGC(state), cell_area.X, cell_area.Y, layout);
}
}
}
Now I want to draw my custom field instead of "Hello, World!".
Some googling suggested SetProperty/GetProperty is what I need, but:
var col = new TreeViewColumn("TreeItem", renderer);
col.AddAttribute(renderer, "tree-item", 0);
Throws an error on runtime since TreeItemCellRenderer has no property named "tree-item".
Also, it is worth noting that SetProperty is not virtual.
So how should I do it?
Actually, the only problem is that defining the value/data to render in your custom cell renderer is not done in the renderer itself, but outside, in a listener method.
In your widget class, where you define the TreeView and its columns and CellRenderers, you can also define a TreeCellDataFunc, which is nothing but a wickedly hidden name for a listener function that is called before a cell is rendered. But its name already has the "Data" inside and that is what you're supposed to do within that function: Supply the data to render.
TreeItemCellRenderer cellRenderer = new TreeItemCellRenderer();
column.SetCellDataFunc(cellRenderer, new Gtk.TreeCellDataFunc(AwesomeDataFunction));
/**
* Supplies the TreeItemCellRenderer with data for the current cell.
* */
private void AwesomeDataFunction(Gtk.TreeViewColumn column, Gtk.CellRenderer cell, Gtk.TreeModel model, Gtk.TreeIter iter)
{
(cell as TreeItemCellRenderer).<some self-declared property> = <any fitting value>;
}
You will likely want use the iter and model to get the data that you want. The actual Render() function of the CellRenderer is called after that function.
All in all, I think that whole TreeView/CellRenderer model is a mess. While it does work and you can do most things with it, it is completely unclear what you are supposed to do to achieve your goals. That one TreeView tutorial is okay, but really only shows the tip of the iceberg. Why no Widgets are used for the cell rendering is beyond me. It would have been far easier.
You just declare a property and use it like in this example.
class CellRendererSimple : CellRenderer
{
//property asigned to by treeviewcolumn
//eg. TreeViewColumn tvc = new TreeViewColumn("Column 0", new CellRendererSimple(),"tree_item",0);
[GLib.Property("tree_item")] //this line seems to be nessasary for the property to be recognised by the treeviewcolumn.
public string tree_item
{
get;
set;
}
protected override void Render(Gdk.Drawable window, Widget widget, Gdk.Rectangle background_area, Gdk.Rectangle cell_area, Gdk.Rectangle expose_area, CellRendererState flags)
{
//this seems to be the minimum to render text
GC gc = widget.Style.TextGC(StateType.Normal);
Pango.Layout layout = new Pango.Layout(widget.PangoContext);
layout.SetText(tree_item);//add the text to thelayout.
//this is the actuall rendering
window.DrawLayout(gc, expose_area.X, expose_area.Y, layout);
}
}
I haven't found an actual good answer, but I did find a workaround.
My goal has always been to display information while keeping a "pointer" to the class that owns that value.
So here's what I did:
Declare only one column (the "label")
Append rows, each with two columns - the label to display and the class instance. That way the first row is displayed but the second is hidden
Now, getting the second column of each row will give me the value I want, without having to deal with custom renderers and without having random data appearing in the tree.
I know this doesn't answer the question, but it does solve my particular problem. In case someone else ends up in this page, I hope this information is still helpful. But if you do find a better solution, feel free to post it anyway.
Related
I'm trying to make a function that positions a label at the right location when the text is translated or changed, so the label wont be in front of a comboBox, textBox etc..
public void PositionControl(Label label, ComboBox comboBox)
{
label.Left = comboBox.Left - label.Size.Width;
}
like this it works with ComboBox, is there a way I can make it work with no matter what object it is for secont var(textBox,label...) in the function.
You may try the below code. It should solve your problem.
Just make sure to pass the correct parameter to it otherwise you can add these codes inside try-catch block and handle the
public void PositionControl(Label label, Control control)
{
label.Left = control.Left - label.Size.Width;
}
This is probably a basic question, but I can't find answers because the terms are generic.
I am building a WinForm aplication. Its purpose is to set up memory in a certain chip. I think the best way to organize the application is to have a user control for each chip type, derived from a generic parent class. Think of the children as "iphone," "android" and "blackberry," derived from a parent class "phone".
VS2017 Designer has a Panel where I want the control to be. On startup, I generate an object of the base class and add it to the panel. When I press a button, the old object is deleted and replaced with a new one. Each class has just one control, a label with distinctive text.
The problem is, after I press the button, I see both texts. The panel's Controls collection has just one element, but I see the text from both objects. I have tried Refresh, Update and Invalidate withe the same results.
What do I have to do to make the old text "go away" so the only thing I see is the latest object?
private ChipMemBase ChipMemControl = new ChipMemBase();
public Form1()
{
InitializeComponent();
//tbFeedback.Text = string.Format(fmtString, 0, 1, 2, 3, 4, 5);
cbChipName.SelectedIndex = 0;
tbVersion.Text = Version;
OriginalWindowColor = tbFeedback.BackColor;
ShowChipMemControl();
PrintToFeedback(Version);
}
private void ShowChipMemControl()
{
var ctl = pnlChipMem.GetChildAtPoint(new Point(5,5));
if (null != ctl)
{
if (ctl != ChipMemControl)
{
pnlChipMem.Controls.Remove(ctl);
ctl.Dispose();
pnlChipMem.Update();
Refresh();
}
}
if (null != ChipMemControl)
{
pnlChipMem.Controls.Add(ChipMemControl);
}
}
private void btnMakeChipMemory_Click(object sender, EventArgs e)
{
ChipMemControl = new ChipMemGen2();
ShowChipMemControl();
}
Screenshots before and after clicking Create
Your ShowChipMemControl gets the control at point 5,5 and checks if it's a ChipMemControl then removes it.
I'm guessing that the reason it's not getting removed is that the control at point 5,5 is not a ChipMemControl.
You can use:
pnlChipMem.Controls.Clear()
to remove all the controls
Or:
ChipMemControl cmc = pnlChipMem.Controls.OfType<ChipMemBase>().FirstOrDefault();
if (cmc != null)
{
pnlChipMem.Controls.Remove(cmc);
cmc.Dispose();
}
To only remove the first instance of ChipMemBase on your pnlChipMem panel.
Got it. The problem was from inheritance, not window behavior. Control lblDefault in the base class, carrying the inconvenient text, was still present in the child class. I had to make it Public in the base class and remove it in the child class constructor:
InitializeComponent();
Controls.Remove(lblDefault);
lblDefault.Dispose();
lblDefault = null;
The clue was this article and project:
dynamically-and-remove-a-user-control
I wrote an Visual Studio 2013 extension that observes Windows.Forms designer windows. When a developer is changing controls in the designer window, the extension tries to verify that the result is consistent with our ui style guidelines. If possible violations are found they are listed in a tool window. This all works fine.
But now I would like to mark the inconsistent controls in the designer window, for example with a red frame or something like this.
Unfortunately, I did not find a way to draw adornments on controls in a designer window. I know that you can draw those adornments if you develop your own ControlDesigner, but I need to do it from "outside" the control's designer. I only have the IDesignerHost from the Dte2.ActiveWindow and can access the Controls and ControlDesigners via that host. I could not find any way to add adornments from "outside" the ControlDesigners.
My workaround for now is to catch the Paint-Events of the controls and try to draw my adornments from there. This doesn't work well for all controls (i.e. ComboBoxes etc), because not all controls let you draw on them. So I had to use their parent control's Paint event. And there are other drawbacks to this solution.
I hope someone can tell me if there is a better way. I'm pretty sure that there has to be one: If you use Menu->View->Tab Order (not sure about the correct english menu title, I'm using a german IDE), you can see that the IDE itself is able to adorn controls (no screenshot because it's my first post on SO), and I'm sure it is not using a work around like me. How does it do that?
I've been googling that for weeks now. Thanks for any help, advice, research starting points....
UPDATE:
Maybe it gets a little bit clearer with this screenshot:
Those blue numbered carets is what Visual Studio shows when selecting Tab order from the View menu. And my question is how this is done by the IDE.
As mentioned I tried to do it in the Paint event of the controls, but e.g. ComboBox doesn't actually support that event. And if I use the parent's Paint event I can only draw "around" the child controls because they are painted after the parent.
I also thought about using reflection on the controls or the ControlDesigners, but am not sure how to hook on the protected OnPaintAdornments method. And I don't think the IDE developers used those "dirty" tricks.
I believe you are seeking for BehaviorService architecture. The architecture with supporting parts like Behavior, Adorner and Glyph and some examples is explained here Behavior Service Overview. For instance
Extending the Design-Time User Interface
The BehaviorService model enables new functionality to be easily layered on an existing designer user interface. New UI remains independent of other previously defined Glyph and Behavior objects. For example, the smart tags on some controls are accessed by a Glyph in the upper-right-hand corner of the control (Smart Tag Glyph).
The smart tag code creates its own Adorner layer and adds Glyph objects to this layer. This keeps the smart tag Glyph objects separate from the selection Glyph objects. The necessary code for adding a new Adorner to the Adorners collection is straightforward.
etc.
Hope that helps.
I finally had the time to implement my solution and want to show it for completeness.
Of course I reduced the code to show only the relevant parts.
1. Obtaining the BehaviorService
This is one of the reasons why I don't like the service locator (anti) pattern. Though reading a lot of articles, I didn't came to my mind that I can obtain a BehaviorService from my IDesignerHost.
I now have something like this data class:
public class DesignerIssuesModel
{
private readonly BehaviorService m_BehaviorService;
private readonly Adorner m_Adorner = new Adorner();
private readonly Dictionary<Control, MyGlyph> m_Glyphs = new Dictionary<Control, MyGlyph>();
public IDesignerHost DesignerHost { get; private set; }
public DesignerIssuesModel(IDesignerHost designerHost)
{
DesignerHost = designerHost;
m_BehaviorService = (BehaviorService)DesignerHost.RootComponent.Site.GetService(typeof(BehaviorService));
m_BehaviorService.Adornders.Add(m_Adorner);
}
public void AddIssue(Control control)
{
if (!m_Glyphs.ContainsKey(control))
{
MyGlyph g = new MyGlyph(m_BehaviorService, control);
m_Glyphs[control] = g;
m_Adorner.Glyphs.Add(g);
}
m_Glyphs[control].Issues += 1;
}
public void RemoveIssue(Control control)
{
if (!m_Glyphs.ContainsKey(control)) return;
MyGlyph g = m_Glyphs[control];
g.Issues -= 1;
if (g.Issues > 0) return;
m_Glyphs.Remove(control);
m_Adorner.Glyphs.Remove(g);
}
}
So I obtain the BehaviorService from the RootComponent of the IDesignerHost and add a new System.Windows.Forms.Design.Behavior.Adorner to it. Then I can use my AddIssue and RemoveIssue methods to add and modify my glyphs to the Adorner.
2. My Glyph implementation
Here is the implementation of MyGlyph, a class inherited from System.Windows.Forms.Design.Behavior.Glyph:
public class MyGlyph : Glyph
{
private readonly BehaviorService m_BehaviorService;
private readonly Control m_Control;
public int Issues { get; set; }
public Control Control { get { return m_Control; } }
public VolkerIssueGlyph(BehaviorService behaviorService, Control control) : base(new MyBehavior())
{
m_Control = control;
m_BehaviorService = behaviorService;
}
public override Rectangle Bounds
{
get
{
Point p = m_BehaviorService.ControlToAdornerWindow(m_Control);
Graphics g = Graphics.FromHwnd(m_Control.Handle);
SizeF size = g.MeasureString(Issues.ToString(), m_Font);
return new Rectangle(p.X + 1, p.Y + m_Control.Height - (int)size.Height - 2, (int)size.Width + 1, (int)size.Height + 1);
}
}
public override Cursor GetHitTest(Point p)
{
return m_Control.Visible && Bounds.Contains(p) ? Cursors.Cross : null;
}
public override void Paint(PaintEventArgs pe)
{
if (!m_Control.Visible) return;
Point topLeft = m_BehaviorService.ControlToAdornerWindow(m_Control);
using (Pen pen = new Pen(Color.Red, 2))
pe.Graphics.DrawRectangle(pen, topLeft.X, topLeft.Y, m_Control.Width, m_Control.Height);
Rectangle bounds = Bounds;
pe.Graphics.FillRectangle(Brushes.Red, bounds);
pe.Graphics.DrawString(Issues.ToString(), m_Font, Brushes.Black, bounds);
}
}
The details of the overrides can be studied in the links posted in the accepted answer.
I draw a red border around (but inside) the control and add a little rectangle containing the number of found issues.
One thing to note is that I check if Control.Visible is true. So I can avoid to draw the adornment when the control is - for example - on a TabPage that is currently not selected.
3. My Behavior implementation
Since the constructor of the Glyph base class needs an instance of a class inherited from Behavior, I needed to create a new class. This can be left empty, but I used it to show a tooltip when the mouse enters the rectangle showing the number of issues:
public class MyBehavior : Behavior
{
private static readonly ToolTip ToolTip = new ToolTip
{
ToolTipTitle = "UI guide line issues found",
ToolTipIcon = ToolTipIcon.Warning
};
public override bool OnMouseEnter(Glyph g)
{
MyGlyph glyph = (MyGlyph)g;
if (!glyph.Control.Visible) return false;
lock(ToolTip)
ToolTip.Show(GetText(glyph), glyph.Control, glyph.Control.PointToClient(Control.MousePosition), 2000);
return true;
}
public override bool OnMouseLeave(Glyph g)
{
lock (ToolTip)
ToolTip.Hide(((MyGlyph)g).Control);
return true;
}
private static string GetText(MyGlyph glyph)
{
return string.Format("{0} has {1} conflicts!", glyph.Control.Name, glyph.Issues);
}
}
The overrides are called when the mouse enters/leaves the Bounds returned by the MyGlyph implementation.
4. Results
Finally I show screenshot of a example result. Since this was done by the real implementation, the tooltip is a little more advanced. The button is misaligned to all the comboboxes, because it's a little too left:
Thanks again to Ivan Stoev for pointing me to the right solution. I hope I could make clear how I implemented it.
Use the System.Drawing.Graphics.FromHwnd method, passing in the HWND for the designer window.
Get the HWND by drilling down into the window handles for visual studio, via pinvoke. Perhaps use tools like Inspect to find window classes and other information that might help you identify the correct (designer) window.
I've written a C# program to get you started here.
I've figured out how to set the tint for the title bar and have set a background image for all my views, but for the life of me I cannot figure out how to set the default UILabel color for section headers and such. I don't want to riddle my code setting all my colors to UIColor.Black by hand. Is there any way to get a list of different UIElements and the way to set defaults (colors, fonts, sizes) for each? I'm specifically interested in the color of labels, but any other information would be extremely helpful for the future.
Starting with iOS 5 there's an UIAppearance class to handle global appearance of many UI elements.
The way it works with MonoTouch is that an inner type, called Appearance, exists on those types. You can directly set the properties from it (easy with MonoDevelop's code completion). E.g.
UILabel.Appearance.BackgroundColor = UIColor.Blue;
Sadly it does not cover everything you might want for every control (nor will it work before iOS 5).
This is kind of hacky, but this does the trick, and it works with all kinds of other UI controls that subclass UIView. This assumes you have your own subclassed DialogViewController (mine has lots of handy time-saving things in it, and all my views subclass my DVC to do work).
protected void ForAllSubviewUIControlsOfType<TUIType>(Action<TUIType, int> actionToPerform) where TUIType : UIView
{
processSubviewUIControlsOfType<TUIType>(this.View.Subviews, actionToPerform, 0);
}
private void processSubviewUIControlsOfType<TUIType>(UIView[] views, Action<TUIType, int> actionToPerform, int depth) where TUIType : UIView
{
if (views == null)
return;
if (actionToPerform == null)
return;
foreach (UIView view in views)
{
if (view.GetType () == typeof(TUIType))
{
actionToPerform((TUIType)view, depth);
}
if (view.Subviews != null)
{
foreach (UIView subview in view.Subviews)
{
processSubviewUIControlsOfType<TUIType>(view.Subviews, actionToPerform, depth + 1);
}
}
}
}
To solve the original problem and change the text colors of all labels that are in sections, this is what you need to do:
public override void ViewDidAppear (bool animated)
{
base.ViewDidAppear (animated);
ForAllSubviewUIControlsOfType<UILabel>((label, depth) => {
if ( /*label.Superview is UIView && label.Superview.Superview is UITableView*/ depth == 1 )
{
label.TextColor = UIColor.Black;
}
});
}
The commented out part is what was determining if the label existed inside a section, before I decided to return an int depth count to see if it was on the view.
EDIT: Seems as if this isn't the perfect solution... the depth count varies for these elements. I'll have to figure out something better in the next day or two.
in my Win Forms app I create an array of dynamic custom controls inside a loop. These, lets call them 'boxes', are like my basic pieces of information. I also create string arrays in other parts of the code that contain the information of this 'boxes', so that for example string[3] is a variable of box[3] and so does stringa[3], stringb[3], stringc[3]... all the arrays with the same index are related to the box with that index. Hope I make myself clear.
Only 2 of this strings are shown in 2 labels inside each custom control 'box' in the array, but the others are there because I want to make something so that when the user clicks one of these controls the other strings can be shown in another control. Sort of something like "More Information...". All the 'boxes' in the array need to have the same event handler because I create +100.
To put it more into context, each custom control 'box' in the array shows the Symbol and the Price of a stock and I want that when the user clicks on each stock more quote information is shown on another special control which is like a placeholder for "More info".
I am thinking of 2 ways to do it:
If I could "detect" the index of the clicked control (which is the same in the strings related to it), I could just set this to an int j and all I have to do is show all the strings a,b,c... with index j. Unfortunately I cannot find a way to do this, maybe it is not even possible.
The other way I have thought is to create some properties for my custom control which "store" this variables, and in my app instead of assigning strings I would set properties for each control, which I could later retrieve when the control is clicked. I haven't tryed this because I don't know exactly how to do it.
What do you think? Do you know how can I achieve this or do you have a different idea that will work? Please help! Thanks in advance.
It's kind of a broad implementation question since there are countless ways you could implement something like this.
If you are creating two collections, one with the buttons and one with the information, you potentially could just assign each of the buttons 'Tag' properties to point to the corresponding info and assign a generic OnClick event handler that displays the info.. something like:
infoControl.text = ((InfoClass)((Button)Sender.Tag)).pieceOfInformation;
But again there are many ways to do this, and the choice comes down to how you store your information.
For your first method, you could have a property of your custom control that is the index.
public class Box : Control
{
// ...existing code
private int index;
public int Index
{
get
{
return index;
}
set
{
index = value;
}
}
}
OR
For your second method, you could have a property of your custom control that is the additional info string.
public class Box : Control
{
// ...existing code
private string extraInfo;
public string ExtraInfo
{
get
{
return extraInfo;
}
set
{
extraInfo = value;
}
}
}
In either case, you could then access the proper information right in your click handler for the "box".
i don't know about the first way - got to noodle around more, but in the second way you can extended your custom or built-in control: for example:
public class ExtendedLabel: Label
{
public string[] MoreInfo { get; set; }
}
and initialize it
public TestForm()
{
InitializeComponent();
ExtendedLabel label = new ExtendedLabel();
label.MoreInfo = new string[] { "test" };
this.Controls.Add(label);
label.AutoSize = true;
label.Location = new System.Drawing.Point(120, 87);
label.Name = "label1";
label.Size = new System.Drawing.Size(35, 13);
label.TabIndex = 0;
label.Text = label.MoreInfo[0];
}
And later in your event handler you can use the inside information