Wonderful Shell routines
-
- Posts: 17
- Joined: Fri Jun 30, 2006 4:26 am
- Location: Germany
Wonderful Shell routines
Hello members,
I'm very happy that I found the wonderful mad Collection, which is very easy to use, well-documented and in my case - for the madShell - the easiest way to handle the often strange Shell objects of the ShellAPI.
Indeed - I worked a lot to understand the concept of the shell routines, including CSIDL, PItemIDLists, etc., but the way the IShellObj interface can be used makes all these things understandable.
The only disadvantage is that the source is not included so I don't know *how* it works exactly, but that's the decision of the author.
Thanks for the great work!
Claus.
I'm very happy that I found the wonderful mad Collection, which is very easy to use, well-documented and in my case - for the madShell - the easiest way to handle the often strange Shell objects of the ShellAPI.
Indeed - I worked a lot to understand the concept of the shell routines, including CSIDL, PItemIDLists, etc., but the way the IShellObj interface can be used makes all these things understandable.
The only disadvantage is that the source is not included so I don't know *how* it works exactly, but that's the decision of the author.
Thanks for the great work!
Claus.
-
- Posts: 17
- Joined: Fri Jun 30, 2006 4:26 am
- Location: Germany
-
- Posts: 17
- Joined: Fri Jun 30, 2006 4:26 am
- Location: Germany
Memory allocation/disallocation when using ShellObj calls
By the way:
when I use the structure
with ShellObj('c:\whatever.txt') do ...
the madshell Unit allocates a new Shell object for performing the action with the shell item 'c:\whatever.txt'. Is the structure correctly dis-allocated when a procedure / the application ends or Do I have to dis-allocate such objects for my own? I'm working on a Treeview routine made with ShellObj calls and I ask myself what about all those allocated shell items?
when I use the structure
with ShellObj('c:\whatever.txt') do ...
the madshell Unit allocates a new Shell object for performing the action with the shell item 'c:\whatever.txt'. Is the structure correctly dis-allocated when a procedure / the application ends or Do I have to dis-allocate such objects for my own? I'm working on a Treeview routine made with ShellObj calls and I ask myself what about all those allocated shell items?
madSecurity is based on reference counted interfaces. This works similar to how Delphi dynamic strings are working. Basically you don't need to care about freeing/destroying. It is done automatically.
I'd suggest that you complete your TreeView project and then afterwards use a tool like MemProof to make sure that there are no memory/resource leaks.
I'd suggest that you complete your TreeView project and then afterwards use a tool like MemProof to make sure that there are no memory/resource leaks.
-
- Posts: 17
- Joined: Fri Jun 30, 2006 4:26 am
- Location: Germany
Hi! You are very fast in replying.
Well, indeed, ShellObj frees the memory pretty much like using temporary strings. That's fascinating! All I have to do is to put anything I want to do with the ShellObj structure within the WITH context. Outside the with structure the object is already freed.
Allow me to publish the code of my testing tree view here (so other can see how easy madShell makes the stuff):
Well, it's only a testing project - nothing real big. It only shows 2 levels of the tree - but you can right-click to get the explorer context menu.
Now I have to take some time to study what madShell definitely performs.
You have very high knowledge of the internal structures of the Shell (but not only), and of course you did that what the ShellAPI programmers did not: Making an understandable structured object for programmers. Thanks for that very good work!
Claus.
Well, indeed, ShellObj frees the memory pretty much like using temporary strings. That's fascinating! All I have to do is to put anything I want to do with the ShellObj structure within the WITH context. Outside the with structure the object is already freed.
Allow me to publish the code of my testing tree view here (so other can see how easy madShell makes the stuff):
Code: Select all
unit Unit1;
interface
uses
Windows, Messages, SysUtils, Variants, Classes, Graphics, Controls, Forms,
Dialogs, ComCtrls, madshell;
type
TForm1 = class(TForm)
TreeView1: TTreeView;
procedure TreeView1AdvancedCustomDrawItem(Sender: TCustomTreeView;
Node: TTreeNode; State: TCustomDrawState; Stage: TCustomDrawStage;
var PaintImages, DefaultDraw: Boolean);
procedure TreeView1MouseUp(Sender: TObject; Button: TMouseButton;
Shift: TShiftState; X, Y: Integer);
procedure FormActivate(Sender: TObject);
procedure FormCreate(Sender: TObject);
private
{ Private-Deklarationen }
first:boolean;
procedure addnode(node:TTreeNode; ishell:IShellObj; var level:integer);
public
{ Public-Deklarationen }
end;
var
Form1: TForm1;
implementation
{$R *.dfm}
procedure TForm1.addnode;
var
i:integer;
begin
inc(level);
if ishell.isfolder then begin
node:=TreeView1.Items.AddChild(node,ishell.name);
if node<>nil then begin
for i:=0 to ishell.ItemCount-1 do begin
application.ProcessMessages;
if level<3 then addnode(node,ishell.Items[i],level);
end; {for}
end; {if}
end; {if}
dec(level);
end;
procedure TForm1.FormCreate(Sender: TObject);
begin
first:=true;
end;
procedure TForm1.FormActivate(Sender: TObject);
var
level:integer;
begin
level:=0;
addnode(nil,shellobj('c:\'),level);
first:=false;
end;
function getpath(node:TTreeNode):string;
begin
result:='';
if node<>nil then begin
result:=node.Text;
while node.parent<>nil do begin
result:=node.Parent.Text+'\'+result;
node:=node.Parent;
end; {while}
result:=stringreplace(result,'\\','\',[rfReplaceAll]);
end; {if}
end;
procedure TForm1.TreeView1MouseUp(Sender: TObject; Button: TMouseButton;
Shift: TShiftState; X, Y: Integer);
var
node:TTreeNode;
path:string;
begin
if button=mbright then begin
node:=TreeView1.GetNodeAt(x,y);
path:=getpath(node);
if path<>'' then shellobj(path).ShowContextMenu(handle);
end; {if}
end;
procedure TForm1.TreeView1AdvancedCustomDrawItem(Sender: TCustomTreeView;
Node: TTreeNode; State: TCustomDrawState; Stage: TCustomDrawStage;
var PaintImages, DefaultDraw: Boolean);
var
rect:TRect;
icon:TIcon;
begin
if node<>nil then begin
rect:=node.DisplayRect(true);
with treeview1.Canvas do begin
brush.Color:=clYellow;
brush.Style:=bsSolid;
rect.Right:=rect.Right+16;
textrect(rect,rect.Left+16+2,rect.Top+2,node.text);
with ShellObj(getpath(node)) do begin
icon:=TIcon.Create;
icon.Handle:=geticon(16);
draw(rect.Left,rect.Top,icon);
icon.Free;
end; {if}
end; {with}
end; {if}
end;
end.
Now I have to take some time to study what madShell definitely performs.
You have very high knowledge of the internal structures of the Shell (but not only), and of course you did that what the ShellAPI programmers did not: Making an understandable structured object for programmers. Thanks for that very good work!
Claus.