AutoCAD DevBlog

Making a non database resident copy of a complex object

By Balaji Ramamoorthy

In ObjectARX you can do a non database-resident copy of a single object using the clone() or copyFrom() methods. But these two methods do not operate on complex entities such as an AcDbBlockReference with AcDbAttributes. As is explained in the documentation, the clone() method calls the copyFrom() method that does a shallow clone/copy of the object. In other words, it does not copy the owned objects (AcDbAttribute). There is no direct method available to do this in ObjectARX.

The solution is to copy each object individually and to reattach the newly created attributes with the block-reference object. Be aware of a special case where the block-definition of the block has constant attributes. These attributes are not sub-entities of the block-reference entity, instead they are AcDbAttributeDefinitions in the block table record (the block definition).

If you want a database resident copy of such complex objects, do not use the deepClone() method directly: you should call the AcDbDatabase::deepCloneObjects() method, ensuring the AcDbObjectId of the block-reference is in the array argument. As deepCloneObjects() follows the AcDbHardOwnerShipId and AcDbHardPointerId relationships, the AcDbAttributes will be copied as well.

Constant and Non-Constant Block Attributes

By Balaji Ramamoorthy

When iterating attributes of a specific block reference, you may not find all of them as expected. The reason for this is because an AcDbBlockTableRecord owns all attribute definitions, and an AcDbBlockReference (the insert of a block) owns all the non-constant attributes. If a block definition defines some constant attributes, AutoCAD will not create attributes in the AcDbBlockReference for them. By iterating the AcDbBlockReference object, you can find only the non-constant attributes, if you also want the constant attributes, you must iterate the AcDbBlockTableRecord as well.

The following sample code demonstrates this:

ads_name ename ;

ads_point pt ;

acedEntSel (L"Select a Block Reference: ", ename, pt) ;

 

AcDbObjectId id ;

acdbGetObjectId (id, ename) ;

 

AcDbBlockReference *pRef ;

acdbOpenAcDbObject ((AcDbObject *&)pRef, id, AcDb::kForRead) ;

 

//----- List Non-Const attributes

AcDbObjectIterator  *pIter ;

pIter = pRef->attributeIterator () ;

int cnt =0 ;

while ( !pIter->done () )

{

    cnt++ ;

    AcDbObjectId ida =pIter->objectId () ;

    acutPrintf (L"\nAttribute %d - ObjectId %ld", cnt, ida) ;

    pIter->step();

}

delete pIter ;

acutPrintf (L"\n%d Non-Const Attribute(s) Found!", cnt) ;

 

//----- List Constant attributes

cnt =0 ;

id =pRef->blockTableRecord () ;

AcDbBlockTableRecord *pRec ;

acdbOpenAcDbObject ((AcDbObject *&)pRec, id, AcDb::kForRead) ;

if ( pRec->hasAttributeDefinitions () == Adesk::kTrue )

{

    AcDbBlockTableRecordIterator *pIter2 ;

    pRec->newIterator (pIter2) ;

    while ( !pIter2->done () )

    {

        AcDbEntity *pEnt ;

        pIter2->getEntity (pEnt, AcDb::kForRead) ;

        if ( pEnt->isKindOf (AcDbAttributeDefinition::desc ()) )

        {

            AcDbAttributeDefinition *pDef=(AcDbAttributeDefinition *)pEnt ;

            if ( pDef->isConstant () == Adesk::kTrue )

            {

                cnt++;

                acutPrintf (L"\nConstant Attribute %d - ObjectId %ld)", cnt, pDef->objectId()) ;

            }

        }

        pEnt->close () ;

        pIter2->step () ;

    }

    delete pIter2 ;

}

acutPrintf (L"\n%d Constant Attribute(s) Found!", cnt) ;

 

pRec->close () ;

pRef->close () ;

Using NOMUTT variable to stop "Command: " Prompt between vla-getXXX methods

By Balaji Ramamoorthy

You may want to suppress the echo of the "Command: " prompt after subsequent calls to any of the 'Getxxxxxx' methods using ActiveX. For this the NOMUTT system variable along with extra Prompts as shown in the following sample lisp code can be used to stop the Command prompt from appearing after each vla-getXXX method.

(defun c:test (/ doc p1 p2 uo)
  (vl-load-com)
  (setq doc (vla-get-ActiveDocument (vlax-get-acad-object))
          msp (vla-get-modelspace doc)
            uo  (vla-get-utility doc)
  )

  (prompt "\nStart point:")
  (setvar "nomutt" 1)
  (vla-InitializeUserInput uo 9)
  (setq p1 (vla-getPoint uo nil))

  (setvar "nomutt" 0)
  (prompt "\nEnd point:")
  (setvar "nomutt" 1)
  (vla-InitializeUserInput uo 9)
  (setq p2 (vla-getPoint uo p1))

  (vla-addLine msp p1 p2)
  (setvar "nomutt" 0)
  (princ)
)

How to detect if a toolbar has been loaded using VLISP?

By Balaji Ramamoorthy

Using VLisp, to detect if a toolbar has been loaded, we can use ActiveX methods to do the work, just like VBA does. The following code can achieve the task:

;;; Detect if the "Web" toolbar is loaded
(defun c:detectToolbar (/ acadapp menuGroups menuGroup0 toolbars)
   ;; Load COM support
   (if (car (atoms-family 1 '("vl-load-com")))
      (vl-load-com)
   )

   ;; Get AutoCAD application, the ACAD menuGroup and toolbars
   (setq acadapp    (vlax-get-acad-object)
           menuGroups (vla-get-menugroups acadapp)
           menuGroup0 (vla-item menuGroups "ACAD")
           toolbars   (vla-get-toolbars menuGroup0)
   )

   (vlax-for eachTB toolbars
      (setq nameTB (vla-get-name eachTB))
      (if (= nameTB "Web")
         (princ "\n Web toolbar has been loaded.")
      )
   )

   (princ)
)

This example determines if a menugroup named "BatchPublish" has been loaded:

(defun c:chkmenu (/ menuGroups eachMG nameMG)
(if (car (atoms-family 1 '("vl-load-com")))
      (vl-load-com)
   )

   ;; Get AutoCAD application, and menuGroups
   (setq acadapp    (vlax-get-acad-object)
           menuGroups (vla-get-menugroups acadapp)
   )

   (vlax-for eachMG menuGroups
      (setq nameMG (vla-get-name eachMG))
      (if (= nameMG "BATCHPUBLISH") ;Change this to the name of your menu
         (princ "\n BATCHPUBLISH menu has been loaded.")
      )
   )
   (princ)
  )

Installing and using RealDWG 32bit on 64bit Windows

by Fenton Webb

If you are developing a RealDWG application for both 32bit and 64bit Windows you have probably wondered if 32bit RealDWG is supported on Windows 64bit…

Let me answer some of your questions.

1. Can you install the RealDWG 32bit SDK on 64bit Windows
   Yes you can, installing the RealDWG 32bit SDK is supported on 64bit Windows. 
   
2. Can I create a RealDWG 32bit application and install it on 64bit Windows.
   Yes you can, but as with all Autodesk products, where we have a 64bit version of the software we don’t support installing the 32bit version on Windows 64bit. The reason is because, although 32bit versions of our software are capable of running on 64bit Windows, due to the huge amount of work needed to test and release a product on a Windows platform, given the total number of user requests for 32bit Autodesk products on 64bit Windows, it doesn’t make business sense to support it.
   
3. Does RealDWG even need to be installed to use it?
   Actually, as long as you don’t use the RealDWG COM API, then you can use it just like you would the ObjectARX SDK – all I do is copy the RealDWG SDK files from machine to machine, I never install it.
   
4. How do I add the RealDWG components to my installer so that they get installed automatically with my runtime?
   It’s really easy; all you have to do is add all of the MSI Merge Modules (MSM) files to your installer MSI. Once you have done that, save the MSI. Next, go to the Property table of the MSI and edit the newly created ODBXHOSTAPPREGROOT property – make sure the value of that property reflects the same value that your RealDWG HostApplicationServices.GetRegistryProductRootKey() returns. Check out the Redistribution Requirements section of the readdbx.chm file found in the RealDWG SDK for more information.

How to check if a given string is a valid name for an item in a symbol table?

By Marat Mirgaleev

Issue

I want to check if a string provided by the user is suitable for a block name or for a layer name. Is there any function that can validate the string?

Solution

You are right, not every name is acceptable, for example, for a block. When creating a block in the AutoCAD UI, the user can get the following error message:

As you may know, there is a number of Symbol Tables in an dwg file, like Block table, Linetype table etc.

To check if a given string is a valid symbol table name or not, in ObjectARX you can use acdbSNValid() function.

AutoCAD .NET API provides the SymbolUtilityServices.ValidateSymbolName() for the same purpose. The method throws an exception when the name is invalid. Here is an example:

string[] names = { "lk3j4!@ #`$>%", // will cause an exception

                   "lk3j4!@ #$%" }; // will work fine

foreach (string s in names)

{

  try

  {

    // Validate the provided symbol table name

    SymbolUtilityServices.ValidateSymbolName( s, false);

 

    System.Windows.Forms.MessageBox.Show( s + " is a valid name." );

  }

  catch

  {

    // An exception has been thrown, indicating that

    //    the name is invalid

    System.Windows.Forms.MessageBox.Show( s +" is an invalid name.");

  }

} // foreach

The 2nd parameter of the ValidateSymbolName() is a flag that allows or disallows the '|' symbol (so called "pipe" symbol).

API for 3DCONFIG

By Adam Nagy

I know I can use the command line version of 3DCONFIG to change graphics settings using e.g. SendStringToExecute(). However, I'm wondering if it's also possible through some direct API.

Solution

Yes, you can access these settings through the AcGsConfig class.

This ARX sample command toggles the Hardware Acceleration setting:

static void ToggleHWAcceleration(void)

{

  AcGsConfig * gsConf =

    acgsGetGsManager()->getGSClassFactory()->getConfig();

  bool b = gsConf->isFeatureEnabled(AcGsConfig::kHwAcceleration);

  gsConf->setFeatureEnabled(AcGsConfig::kHwAcceleration, !b);

  gsConf->saveSettings();

}

The same in .NET:

[CommandMethod("ToggleHWAcceleration")]

public void ToggleHWAcceleration()

{

  using (Autodesk.AutoCAD.GraphicsSystem.Configuration config =

    new Autodesk.AutoCAD.GraphicsSystem.Configuration())

  {

    bool b = config.IsFeatureEnabled(

      Autodesk.AutoCAD.GraphicsSystem.

      HardwareFeature.HardwareAcceleration);

    config.SetFeatureEnabled(

      Autodesk.AutoCAD.GraphicsSystem.

      HardwareFeature.HardwareAcceleration, !b);

    config.SaveSettings();

  }

}

PInvokeStackImbalance error

By Adam Nagy

I'm trying to P/Invoke acedPostCommand() so that I could use it to cancel the current command, but I get a PInvokeStackImbalance error when calling acedPostCommand inside my .NET code:

How could I solve this?

Here is the P/Invoke I'm using:

[DllImport("acad.exe",

  CharSet = CharSet.Unicode,

  EntryPoint = "?acedPostCommand@@YAHPB_W@Z")]

public static extern bool acedPostCommand(string cmd);

Solution

You also need to specify the calling convention. Once that's done acedPostCommand() works fine:

[DllImport("acad.exe",

  CharSet = CharSet.Unicode,

  CallingConvention = CallingConvention.Cdecl,

  EntryPoint = "?acedPostCommand@@YAHPB_W@Z")]

public static extern bool acedPostCommand(string cmd);

 

Start command with escape characters ^C^C

By Adam Nagy

I'd like to start some commands from my modeless dialog's buttons. I already found that in case of SendStringToExecute I should use '\x03' characters instead of ^C. My only problem is that in this case if no command is running then I end up with two *Cancel* strings in the Command Window, whereas a menu's ^C^C does not cause *Cancel* to appear if no command is running when the menu item is clicked.

How could I achieve the same?

Solution

You could check the CMDNAMES system variable and depending on how many commands are currently running you could add that many escape characters to the beginning of your command string.

Here is a sample that demonstrates this:

using System;

using System.Windows.Forms;

using acApp = Autodesk.AutoCAD.ApplicationServices.Application;

using Autodesk.AutoCAD.ApplicationServices;

 

namespace CsMgdAcad1

{

  public partial class Form1 : Form

  {

    public Form1()

    {

      InitializeComponent();

    }

 

    private void button1_Click(object sender, EventArgs e)

    {

      string esc = "";

 

      string cmds = (string)acApp.GetSystemVariable("CMDNAMES");

      if (cmds.Length > 0)

      {

        int cmdNum = cmds.Split(new char[] { '\'' }).Length;

 

        for (int i = 0; i < cmdNum; i++)

          esc += '\x03';

      }

 

      Document doc = acApp.DocumentManager.MdiActiveDocument;

      doc.SendStringToExecute(esc + "_.LINE ", true, false, true);

    }

  }

}

Creating block from selection set

By Virupaksha Aithal

Below code shows the procedure to create a new block table record from entities selected on screen.  The important function used is “DeepCloneObjects” which takes collections of object ids to be copy. The second parameter is the “owner” of the copied entities. In the below code, the block table record (“test”) is passed as owner, so that the selected entities clone objects are copied to block table record.

[CommandMethod("CreateBlock")]

public void CreateBlock()

{

    Document doc = Application.DocumentManager.MdiActiveDocument;

    Database db = doc.Database;

    Editor ed = doc.Editor;

 

    TypedValue[] filterlist = new TypedValue[1];

    filterlist[0] = new TypedValue(0, "CIRCLE,LINE");

 

    SelectionFilter filter = new SelectionFilter(filterlist);

 

    PromptSelectionOptions opts = new PromptSelectionOptions();

    opts.MessageForAdding = "Select entities: ";

    PromptSelectionResult selRes = ed.GetSelection(opts, filter);

 

    if (selRes.Status != PromptStatus.OK)

    {

        return;

    }

 

    ObjectId[] ids = selRes.Value.GetObjectIds();

    ObjectId blockId = ObjectId.Null;

    //add a block table record with name test...

    using (Transaction Tx = db.TransactionManager.StartTransaction())

    {

        BlockTable bt = (BlockTable)Tx.GetObject(db.BlockTableId,

                                                   OpenMode.ForRead);

 

        if (!bt.Has("Test"))

        {

            bt.UpgradeOpen();

            //create new

            BlockTableRecord record = new BlockTableRecord();

            record.Name = "Test";

 

            bt.Add(record);

            Tx.AddNewlyCreatedDBObject(record, true);

        }

 

        blockId = bt["Test"];

        Tx.Commit();

    }

 

    //copy the select entities to block by using deepclone.

    ObjectIdCollection collection = new ObjectIdCollection(ids);

    IdMapping mapping = new IdMapping();

    db.DeepCloneObjects(collection, blockId, mapping, false);

 

}