CHSPACE command altering the Z-values of objects to -#.##e99

[hencoop]

The CHSPACE command is jacking up Z-values of objects.  If they are MULTILEADERs, the Z-values CANNOT be fully reset to 0.0 using the CHANGE command, Autolisp, or Visual Lisp.  There is at least one unreachable Z-value that is untouched.  It can be seen using the LIST command on the MULTILEADER (the ASSOC 10, 12, 110 values have Z=0.0).

MULTILEADER Layer: "C-VI046NOTE" Space: Model space Handle = 6e0d Leader Type: StraightLeader Content Type: MTextContent 
Landing: ON Leader Number: 0
Vertex: X = 1643382.1137, Y = 581080.0480, Z = 0.0000 
Vertex: X = 1643392.4306, Y = 581103.7736, Z = -6.3281E+99
Multileader style: STandard Annotative: NO

I exploded one of these problem MULTILEADERs which had spline leaders (to recreate it from the text) and the splines disappeared; however, after saving the drawing, closing it, the drawing immediately crashed upon reopening.  I ran RECOVER on it and that found the splines which had to be deleted to fix the drawing.

Any ideas why CHSPACE is screwing the Z-values?

[Patchy]

It has the same value as BACKZ, but it's read only and no one knows why it happens or how to fix it.

Only happened with Land Desktop and C3D.

[hencoop]

My drawing shows BACKZ = 0.0000 (read only)

[hencoop]

I checked all of the System Variables and these are the only two that have large Z-values:

EXTMAX 34.0005,22.0003,xxx-xxxxxxxx.0000 (read only)
EXTMIN -0.0005,-0.0003,-1.5258E+15 (read only)

[Patchy]

Activate the viewport then use command BACKZ

[hencoop]

There it is!

BACKZ = 6.3281E+99 (read only)

[cadffm]

The formula is  Z of your object * Viewportscale = New Z in paperspace.

[hencoop]

I rechecked the System Variables with the viewport activated.  The Z-value of the TARGET seems to be the source of my problems.

These were the prior values (idk why my previous reply x'd out the EXTMAX Z-value numbers):
EXTMAX            34.0005,22.0003,xxx-xxxxxxxx.0000     (read only)
EXTMIN            -0.0005,-0.0003,-1.5258E+15          (read only)

Viewport Activated:
EXTMAX            3.4191E+99,585628.2073,942.0457      (read only)
EXTMIN            -1.7315E+84,-5.3249E+99,-1.2656E+99  (read only)
TARGET            1643490.8475,581068.6715,-6.3281E+99 (read only)

[hencoop]

The Z-value of the objects are all 0.0000

[Patchy]

It doesn't matter what viewport scale, it will inherit the BACKZ value when chspace is used.

Here is the dwg, you can give it a test.

[cadffm]

@Patchy  In case the objetcs are not at 0 (isn't the case here in this thread)- you are wrong.

[hencoop]

Something is really screwed in this drawing.

I tried to reset everything by setting PLAN to WORLD and then using the DVIEW command to reset the twist and ZOOM to reset the scale.  The magnification height would not reset from 7.1997E+99 using ZOOM 1/50XP or entering a magnification height of 1000.  It looks like the only cure for this will be to WBLOCK Mspace and Pspace and then open the Mspace wblocked drawing and insert the Pspace wblocked drawing into its layout and explode it.  This should create a new drawing header and leave out the corruption that is causing this... we'll see.

Command: (setq itc (getvar "viewctr"))
(1.64351e+06 581049.0 0.0)

Command: (setq vangl (getvar "viewtwist"))
0.0601247

Command: PLAN

Enter an option [Current ucs/Ucs/World] <Current>: w
Regenerating model.

Command: (SETVAR "viewtwst" vangl)

ERROR: AutoCAD variable setting rejected: "viewtwst" 0.0601247; reset after error

Command: VIEWTWIST

VIEWTWIST = 0.00000000 (read only)
Command: (* 180.0 (/ vangl PI))
3.44489

Command: (* 180.0 (/ vangl PI))*Cancel*

Command: *Cancel*

Command: VIEWTWIST

VIEWTWIST = 0.00000000 (read only)
Command: DVIEW

Select objects or <use DVIEWBLOCK>:

Enter option
[CAmera/TArget/Distance/POints/PAn/Zoom/TWist/CLip/Hide/Off/Undo]: tw

Specify view twist angle <0.00>: (* 180.0 (/ vangl PI))
3.44489

Enter option
[CAmera/TArget/Distance/POints/PAn/Zoom/TWist/CLip/Hide/Off/Undo]:
Regenerating model.

Command: Z
ZOOM
Specify corner of window, enter a scale factor (nX or nXP), or
[All/Center/Dynamic/Extents/Previous/Scale/Window/Object] <real time>: c

Specify center point: !itc
(1.64351e+06 581049.0 0.0)

Enter magnification or height <7.1997E+99>: 1000
Regenerating model.

Command:
ZOOM

Specify corner of window, enter a scale factor (nX or nXP), or
[All/Center/Dynamic/Extents/Previous/Scale/Window/Object] <real time>: *Cancel*

Command: *Cancel*

Command: Z
ZOOM
Specify corner of window, enter a scale factor (nX or nXP), or
[All/Center/Dynamic/Extents/Previous/Scale/Window/Object] <real time>: 1/50xp
Regenerating model.

[hencoop]

Not wrong, just misunderstood.

[cadffm]

@hencoop 

then you found it, target not perpendicular to object plane,

this way your statement "all at Z=0" is right, if you talking about WCS or UCS,

but the data isn't "all at Z=0" messured from the view plane.

Next time, upload a sample dwg - it's much better for other users and a (shorter) discussion 🙂

[hencoop]

@Patchy , @cadffm 

I recreated the drawing using WBLOCK as described in my prior reply.  Now, all is is good!  Something CHSPACE jacked up the drawing!  

New drawing results:

    Command: MS
    MSPACE
    Command: TARGET
    TARGET = 1642771.3801,583175.9174,341.5070 (read only)
    Command: BACKZ
    BACKZ = 0.0000 (read only)

New drawing results with no other edits except one use of CHSPACE!

    Command: CHSPACE

    Select objects: 1 found

    Select objects:
    Set the SOURCE viewport active and press ENTER to continue.:
    1 object(s) changed from MODEL space to PAPER space.
    Objects were scaled by a factor of 0.02 to maintain visual appearance.

    Command: Specify opposite corner or [Fence/WPolygon/CPolygon]:
    Command: MS
    MSPACE
    Command: BACKZ

    BACKZ = 6.3281E+99 (read only)
    Command: TARGET

    TARGET = 1643490.8475,581068.6715,-6.3281E+99 (read only)

😳

My solution?  NEVER USE CHSPACE AGAIN! No, wait!  I need to replace the viewports with new ones and see if it is a viewport corruption problem!

[hencoop]

@cadffm Maybe I don't understand you?  

I am using WORLD coordinates; I have used the command PLAN and set the viewport to WORLD;  Using DVIEW to check the target I see the angle from the XY plane is -90, and, the angle in the XY plane from the X axis is -90, The system variable AUPREC=8

I have just proven to my satisfaction that the cause of this problem is the DVIEW command.  Specifically, the following will jack up/down the values of TARGET and BACKZ:

(COMMAND "._dview" "" "tw" 3.44489 "").  I have a routine that calculates the desired angle and feeds it as a variable name instead of 3.44489 (or 3.444889250415612, the calc'd value to 16 places)

I tested this, it makes no difference whether each individual entry is manually keyed in at the command line or entered by an autolisp function as shown (whether angle is rounded to 4-places or not rounded at all).

What is [wrong with/designed into] DVIEW TWIST that it jacks the TARGET and BACKZ system variables?

Apparently, I cannot twist my viewport to this angle (and likely many other angles) without jacking TARGET and BACKZ and causing these jacked values to be assigned to objects operated on by CHSPACE.

Am I using DVIEW TWIST incorrectly?

[hencoop]

@cadffm, @Patchy 

It seems that this problem may be caused by DVIEW TArget and CAmera points being very close together in Z-value.  (I still don't know what cause that).  Trying to set them coincident to test failed as AutoCAD does not allow them to be coincident... so I set them very close to each other but that did not cause DVIEW TWist to recreate the problem.  Nevertheless, there were similarities between this problem and another I had some while back:

https://forums.autodesk.com/t5/civil-3d-forum/cannot-select-3d-lines/m-p/6223167#M297306 

I've modified my code to set both FRONTZ and BACKZ, turn off clipping, and reset TARGET Z-values (POints option sets this) before my DVIEW TWist command.  That seems to have solved my problem... so far.

Summary:  DVIEW TWist sometimes changes the TARGET and/or BACKZ Z-values to huge or minuscule numbers (#.#e+99).  It seems especially so if they are already not zero.  Resetting FRONTZ, BACKZ,  turning OFF DVIEW CLipping and resetting the TARGET (POints option) before executing DVIEW TWist keeps these Z-values from being jacked... so far.

;Use (GETVAR "VIEWCTR") to keep the view in the same place after DVIEW 
(COMMAND "._DVIEW" "" "CL" "B" "0.0" "CL" "F" "0.0" "CL" "OFF" "PO" (LIST (CAR (GETVAR "VIEWCTR"))(CADR(GETVAR "VIEWCTR"))0.0) (LIST (CAR (GETVAR "VIEWCTR"))(CADR(GETVAR "VIEWCTR"))1.0) "TW" vwang "")

[hencoop · Accepted solution]

@Patchy @cadffm  So, today my fix was failing because AutoCAD saw TARGET and CAMERA POints as coincident.    I was trying to set them to the VIEWCTR X,Y and give them differing Zs.  It appears that their existing Z-values  being minuscule/enormous are responsible for this error in AutoCAD.  The fix is as follows:

To be very clear, all of these DVIEW things must be done before using CHSPACE whenever my viewport's TARGET is not well behaved , i.e. When TARGET has a minuscule/enormous Z-value like -6.385E+99.

Once TARGET is 0,0,0, BACKZ=0.0, and FRONTZ=0.0 using CHSPACE does NOT jack the Z-values of the objects!

This is what was happening (Text Window command history):

Command: DVIEW

Select objects or <use DVIEWBLOCK>:

Enter option
[CAmera/TArget/Distance/POints/PAn/Zoom/TWist/CLip/Hide/Off/Undo]: PO

Specify target point <1643253.5316, 583856.1333, -4.6094E+99>: (LIST(CAR(GETVAR "VIEWCTR"))(CADR(GETVAR "VIEWCTR"))0.0)
(1.64325e+06 583856.0 0.0)

Specify camera point <1643253.5316, 583856.1333, -4.6094E+99>: (LIST(CAR(GETVAR "VIEWCTR"))(CADR(GETVAR "VIEWCTR"))1.0)
(1.64325e+06 583856.0 1.0)

Camera and Target may not be coincident

Here is what I had to do to make it all work for me:

It was necessary to separate the CLipping part of the command from the POints and TWist parts; then, set the TARGET and CAMERA POints to 0,0,0 and 0,0,1 respectively to avoid the coincident error that AutoCAD is reporting; then, set the TWist angle; finally, ZOOM Center using the saved VIEWCTR point.  This works for me... so far.  the results are BACKZ=0.0; FRONTZ=0.0; TARGET= #.###, #.###, 0.0 (and VIEWCTR as I need it)👍

I.e.

      (SETQ vwctr (LIST (CAR (GETVAR "VIEWCTR"))(CADR (GETVAR "VIEWCTR")) 0.0))
      (COMMAND "._DVIEW" "" "CL" "B" "0.0" "CL" "F" "0.0" "CL" "OFF" "")
      (COMMAND "._DVIEW" "" "PO" "0,0,0" "0,0,1" "")
      (COMMAND "._DVIEW" "" "TW" vwang "")
      (COMMAND "._ZOOM" "C" vwctr "")

P.S. The above code is a part of my tool for setting any VIEWTWIST required in my viewport.  It seemed like the most logical place to force these SYSVAR resets.

[Patchy]

Strange that it happened only with Civil drawings.

[cadffm]

😔