QCAD Bugtracker

Andrew,

For a selection of 100 entities with only 9 polylines the tool reports:
“Inserted nodes in self-intersections in 100 polylines”

None of those 9 polylines were self-intersecting.

insertVerticesAtSelfIntersections returns a boolean.
Presumably true/false on success, on change.

Then:
- If false the polyline must not be added to the operation to update it.
And it should not be counted as one that is updated.
- If true the polyline must be updated and counter num must be increased.
- If not a polyline entity it is skipped and should not be counted.

To be sure one can always compare the number of vertices before and after.

Probably an adapted copy of PolylineNormalize.js that has a similar false report.
“100 polylines normalized” Some references to this script are still trailing.

Again one could compare the number of vertices before and after.

Regards,
CVH

Andrew,

Initial tests for this reply on the forum failed because of a space.
I tend to add a space after a comma by habit.

For QCAD Pro RColor::init() additionally loads the user palette.
A list of colorstrings separated by an EOL and split up per line.

An empty colorstring is skipped (e.g. first of every imported palette)
Each of the colorstrings is split into tuples at a comma.

Not at least 2 tuples adds a separator line.
The last tuple is considered as a color code.
This is not validated as in range ‘#000000’-’#ffffff’, starting with ‘#’ + 6 characters ‘0-f’.

Anything before the last comma is considered as the custom name or ‘title’.
This may be empty, include leading/trailing whitespace and/or include a comma.

Before the first custom entry it adds a separator line.

A color code starting with a space would be invalid.
Trailing spaces are not distinguishable in the preference page and also invalid.

It would already be neater to init a custom color from the list with:

init(title.trimmed(), RColor(code.trimmed()));

Presuming that .trimmed() is the cpp analog for String.prototype.trim in JS.

Regards,
CVH

Adding/inserting additional layout blocks in an order that makes sense to the user throws out system numbering out of order...
Unable to sort Layouts in alphanumerical order like other blocks.
Displaying Layout order number in the layout name is pointless since system cannot cope with non-chronological layouts.
Extract from the thread on the forum (linked below)

Finally traced back to a [url=https://github.com/qcad/qcad/blob/9965f5c6c29cc7419d26d37ebffe119d7511641b/src/core/RStorageBlockSort.cpp]sorting function[/url] called in line 247 (Current state):
- model space always first
- layouts always before other blocks
- sort among layouts by tab order
- sort by layout name or block name (ignoring * at start)
 :!: Notice that the last rule will never be reached when comparing 2 Layout objects with the same 'tab order'.
The third rule will return "false" anyway and things are left unsorted as they are queried from the document.

Meaning that we can force any listing order but only by editing the list order values.

Andrew,

When in RCircle::getTangents(point) the Thales circle radius is less than the radius of the given circle then the point is inside the circle resulting in: No tangents.

But when the point is on or very near the given circle the resource defines a tangent that connects the point and the single intersection point with the touching Thales circle.
The result may be erratic depending the outcome of getIntersectionPointsCC.

A tangent from a point to a circle or any limited line segment that is equal or shorter than 1e-6 has no well defined orientation: RLine::getAngle() returns zero by default.

The correct single tangent orientation would be orthogonal with a line towards the center of the given circle. Not zero by default or inline with the normal at that point.

Also excluding touching circles is thus mathematically more correct.

    if (thalesRadius <= getRadius()/2.0) {
        return ret;
    }

Introducing a tolerance is tricky. RS.PointTolerance = 1e-9 for example.
Because for a large circle there can be 2 tangents longer than 1e-6 to a point that is not further than 1e-9 outside the circle.
And because that getIntersectionPointsCC exploits a dynamic tolerance related to circle radii for other reasons.

Workaround: A valid point outside a circle has 2 (limited) tangents with the circle.
None or just one RLine shape can then be excluded as a situation with valid tangents.

Counter indication found in Draw/Line/LineTangent1/LineTangent1.js and RSnapTangential::snapEntity(...) for a circle or an arc where a single solution is considered as valid.

Regards,
CVH

Andrew,

User thhz reports that when starting QCAD with the -always-load-scripts switch a template is not loaded. Instead a drawing is created with only layer 0 ...
... Typically a plain new file.

Besides that, it seems that there are just a minimal count of Linetypes available in such a drawing document: ‘By Layer’, ‘By Block’ and ‘Continuous’.

NewFromTemplate.js is a Pro Addon.
Unstated if other Pro features are also affected.

Reported on any Win10 or Win11 64bit system he tested this on.
Reported for older installations: QCAD 3.27.?? ; QCAD/CAM 3.30.?? ; QCAD/CAM 3.26.4

It is not an issue of a crippled custom script.
On my advice also replicated with the latest trial versions without including any custom scripts at all. A clean installation fails thus with using the switch.

I can not replicate this on a Win7 32bit system with any QCAD installation that I have.
The issue popped up with the script I wrote to access XP preferences in direct for him.
The real fix for this little helpful tool was on another level. My mistake.

He wasn’t aware of that specific start switch and restarted QCAD each edit of a custom script. Exploiting that for editing on the fly.
But now it seems that templates are not loaded ... ?!

The advice is then to not include the switch for production use.

Regards,
CVH

Andrew,

For a test related to: https://www.qcad.org/rsforum/viewtopic.php?f=32&t=10981#p45436
I copied a bitmap as PNG to my custom library path in the user data location.

The library Browser previews the PNG file but it fails to import the bitmap with this warning on the Command History:

'Cannot import file from URL: file:///C:/Users/Admin/AppData/Local/QCAD/QCAD/libraries/Custom/xxxxxx.png'

As far as I can see is the path perfectly OK except the “file:///C:” part.

I get a similar warning when using the default library path under my installation folder.
Functional with for example a PDF file.

I can insert the bitmap with IM.
→ Last used file location = “C:/Users/Admin/AppData/...” (QCAD.ini)
And I can import it (Ctrl+Shift+I).
→ Last used file location = “C:/Users/Admin/AppData/...” (QCAD.ini)

Regards,
CVH

Andrew,

To replicate:
- Start Circle2TP with a few appropriate entities in the drawing
- Indicate 2 entities
- Indicate a point (pos3) ⇒ Previews all solutions
- Trigger an escape event
- Move to near another second shape

Before even picking the other second shape ...
... all possible solutions with former pos3 are displayed.
After picking the other entity the solutions ‘jump’ to solutions with the current position.
The current position will be rather near entity 2 than near the last picked position.
Not critical, just annoyed

Reason:
this.pos3 is not cleared in Circle2TP.State.ChoosingShape2 and Circle2TP.State.ChoosingShape1

It is thus retained after an escape event setting state Circle2TP.State.ChoosingShape2

this.pos1 and this.pos2 are retained too after escape events but serve no further purpose.

Regards
CVH

Andrew,

Related to: FS#2488

Would it be possible to support storing/loading custom property titles starting with “QCAD...” and then also treated as key-value pair.
They don’t need to be supported in the GUI, the Property Editor or elsewhere, storing and reloading them as defined in this session is perfectly fine.

As far as I know it is the only effective way to hide custom properties in the PE, avoiding that users can see specific properties per entity or edit them.

This is also effective:

    var att = new RPropertyAttributes();
    att.setReadOnly(true);
    att.setInvisible(true);
    RObject.setCustomPropertyAttributes("QCAD", "<AnyRequired>", att);

But only in QCAD sessions where this code has been executed.

For some things I can also exploit custom properties of Blocks or Layers.
And those are fairly ‘hidden’ for most users too.
But for my use it’s all about fully immutable and per entity.

Regards,
CVH

Andrew,
https://qcad.org/doc/qcad/3.0/developer/class_duplicate.html
The description is: ‘Copies the current selection to the clipboard ...’

Can be sourced back to:
https://github.com/qcad/qcad/blob/master/scripts/Edit/Duplicate/Duplicate.js#L24-L25
Where the same is stated as in:
https://github.com/qcad/qcad/blob/master/scripts/Edit/Copy/Copy.js#L24-L25

Regards,
CVH

Related topic: https://www.qcad.org/rsforum/viewtopic.php?f=33&t=10901#p45052

The glyph for 0×0022 and 0x003b are indeed identical.

Dug up a version from 2012, all lines, different header type, but that also resembles a semicolon.

Regards,
CVH

Andrew,

The documentation list RLayer::Id while it are RLayerState::Id that are returned.
In essence both are object IDs.
It is correct as in the associated remark: ‘Set of layer state IDs.’

The typo can be sourced back to RDocument.cpp:
https://github.com/qcad/qcad/blob/master/src/core/RDocument.cpp#L1345

QSet<RLayer::Id> RDocument::queryAllLayerStates() const {

Regards,
CVH

Andrew,

Related topic: https://www.qcad.org/rsforum/viewtopic.php?f=30&t=10801

It is stated that:
RLeaderData Class documentation was generated from the following files:
- RLeaderData.h
- RLeaderData.cpp

dimaszOverride() and dimScaleOverride() are deprecated in the current content of:
RLeaderData.h
See: https://github.com/qcad/qcad/blob/master/src/entity/RLeaderData.h RLeaderData.cpp
See: https://github.com/qcad/qcad/blob/master/src/entity/RLeaderData.cpp

The change was committed on Nov 4, 2021:
https://github.com/qcad/qcad/commit/10a424767edbe6b926e94da80d52836fb1341191

It is unchecked if this is a stand alone exception.

Regards,
CVH

Andrew,

Probably not OS or QCAD version related.

The ellipse arc in the example file is drawn CCW and then the start parameter was set to 180. (All part of a test of newer code for the nearest point on an ellipse)
Its end parameter was originally 44.56263613...

If we reload this file its properties are:
- Not reversed (CCW)
- ratio is 0.42307692...
- Parameters are 180 & 404.56263613...
- The reported length is about 1993.21829743...

To replicate the issue you must draw over this ellipse shape:
- EA ; Direction CCW
- SR for snapping to references.
- Indicate the major, minor, start and end point all by snapping to the original references.

You now have a copy of the provided ellipse shape.
Select it with Alt-Select, it is the one without a length.

Properties are:
- Not reversed (CCW)
- ratio is 0.42307692...
- Parameters are 180 & 44.56263613...
- The reported length is NaN

I can replicate that for some other CCW ellipse arcs ...
... Perhaps given that the start parameter is 180 degrees.
Also suspecting that it can be replicated for CW arcs.

For the record:
A 2D wire Centroid for this shape also fails.
Simply because it tests for the length reported by QCAD.
It doesn’t start to approximate something that is about zero long.

Regards,
CVH

Andrew,

I know that GCodeMM.js and GCodeIN.js are generic postprocessors.
Mostly used as base for other or custom postprocessor.

The distinct nature is pretty clear from the filename and the suffix in the display name.
Another distinct nature is that they include this.unit = RS.Millimeter / RS.Inch intended for the CAM side.

Why don’t they include at least a G20/21 in a generic header?

There are several more specific postprocessors where I miss these G-codes for the measurement system.

True, if they are not supported, no error messages would occur when omitted.
The danger is that users rely on build in postprocessors hoping that it matches with the default measurement system of their setup.

IMHO it is safer to trigger an instruction fault before attempting to execute a wrong motion.
As fail-save as possible should always be the goal.

Regards,
CVH

Andrew,

The _DatumFilled Block used by Datum entities (DT) is considered as not used and is purged by Purge Unused Blocks (BP).
There is no block association for these special types of arrowheads in QCAD.

After purging the Datums are not immediately updated but the special arrowheads will eventually disappear. Adding new Datums will not fix that.

Not problematic because the _DatumFilled Block is re-created on save/reload when Datum entities exists. Datums are updated.

Probably not OS or version related.
Related forum topic: https://www.qcad.org/rsforum/viewtopic.php?f=89&t=10694

Best solution would be to support different types of arrowheads and all related:
https://qcad.org/bugtracker/index.php?do=details&task_id=695 https://qcad.org/bugtracker/index.php?do=details&task_id=757 https://qcad.org/bugtracker/index.php?do=details&task_id=1157

Regards,
CVH

See:
https://www.qcad.org/rsforum/viewtopic.php?f=31&t=10379

Could you add “plain text paste” as ctrl-shift-V
Why? Because sometimes the clipboard comes in with rich text formatting that’s anti-helpful.
The workaround is to paste into a plain text editor, or the console, then recopy and repaste into qCad.

Either way the ctrl-shift-V shortcut is useful for those trained on other software but:
one could also argue that qCad should NEVER take rich text, and rather only accept the plain text with both CTRL-V and CTRL-SHIFT-V.

— CHV says

Agreed, having encountered it numerous times while documenting with text snippets from elsewhere. :roll:
Even copying Info measurements from the Command Line are Rich Text.

The best place to file a feature request is QCAD Bugtracker:
https://www.ribbonsoft.com/bugtracker/ You may need a different account there or make a new one.

May I remark that Ctrl-Shift-V is already in use for Paste along Entity (PE)
But not when the Text Editor Widget has the focus and dialog Widgets are always modal on a Win OS system. :wink:
Still, I’ll vote for it in any way.:P

Regards,
CVH

Andrew,

One Microinch is equivalent to 25.4 nanometers and I don’t consider that to be metric.

RUnit.isMetric(this.getDocument().getUnit()) returns true for RS.Microinch or value 8.

https://github.com/qcad/qcad/blob/master/src/core/RUnit.cpp#L145-L155
Returns false for: RS::Inch ; RS::Foot ; RS::Mile ; RS::Mil ; RS::Yard
... and true for all the rest.

Also see last note in: https://www.qcad.org/rsforum/viewtopic.php?f=31&t=9506&p=38295

Not to be confused with RDocument::isMetric() because that returns the measurement system unless that is RS::UnknownMeasurement what is rather impossible with drawings created by QCAD.

Regards,
CVH

Andrew,

A user essentially pointed out to an ACAD textbook example.
https://qcad.org/bugtracker/index.php?do=details&task_id=2533
Step 3 of: https://autodesk.blogs.com/between_the_lines/2012/07/drawing-with-surveyors-units-in-autocad.html

Please refer to this topic:
https://www.qcad.org/rsforum/viewtopic.php?f=33&t=10652

Regards,
CVH

Andrew,
On behalf of user Kmeelo

Polygonal Area adds measurements text with 12 decimal digits and then as short as possible.

Current formatting path see: https://www.qcad.org/rsforum/viewtopic.php?f=32&t=10450&p=42833#p42805

Related forum feature request: https://www.qcad.org/rsforum/viewtopic.php?f=31&t=10458

I also see a misconception of the term ‘precision’.
Numbers precision is the amount of significant digits disregarding the position of the decimal point.
12 decimal digits is about the largest precision for numbers with 5 integer digits.

Regards,
CVH

Suggestion to add spelling checker / spell checker for all text boxes

Please view the attachment.

Regards

Graeme

I see the following problem:
When I open a particular DXF file in the Professional Trial version 3.27.9.0 on macOS, “^I” in text fields is expanded to whitespace.
When I open the same DXF file in the Community Edition 3.27.9.0 on NetBSD/x86_64 (installed from pkgsrc), “^I” in text fields is visible as “^I”.

Andrew,

Most probably not related to OS or QCAD version.

When hatching an area with Hatch from Segments (HS) there might be false positive warnings on open loops.
It seems that this is related to the orientation of the indicated segments and the order of selection.
The real problematic position is not always the first that is reported.
Loop count is increased even if it is a single loop.

Included a basic test drawing to replicate these.
Further details to replicate the issue are in the file.

Regards,
CVH

Andrew,

I can’t find any documentation concerning OffsetBothSides under Misc .. Modify.
It is not listed in the QCAD features nor in the changelog.

There was a Feature Request FS#1130 for this, I think you might want to close that as being implemented.

The tool asks for a Line, Arc, Circle, Ellipse, Spline or Polyline.
It also seems to act on shapes that can be exploded to such entities like Text, Dimension and Hatch entities.

There may be many ‘Result loop not closed’ or ‘No offsets possible’ warnings when hovering over entities on a drawing.
The preview may be partially, one or two sided and even two sided partially and closed with semi-circles.

I only successfully created a dual offset for open Splines, open or logically open Polylines.
All other attempts result in a single offset at the indicated side.

There are the usual exceptions for Polylines with arc-segments but that is related to small arc segments and being tangentially connected.
Only tested it on a Win7 platform.
I don’t suspect that version 3.27.8 includes a fix for that ...

Regards,
CVH

Minor issue!
When using keyboard to initiate the dropdown menus.
I noticed the ALT “D” does not cycle through the headings Draw & Dimension in the same way.
ALT “M” cycles though Modify & Misc or ALT “S” cycles though Select & Snap.
User Name: StanTibbs
E-mail: [email protected]

Windows 7 pro 32 bit msi-Installer
Version: 3.26.4.5 (3.26.4.5)
Internet: QCAD.org
Build Date: Jul 20 2021
Revision: e596276
Qt Version: 5.13.2
Architecture: i386
Compiler: Unknown

Andrew,

In the attached dxf:
Besides that it is clear that duplicate vectors are not filtered ...
Where a segment is crossing in another node a non valid node is returned as self-intersection point.
Seems to be unrelated to CW/CCW or tolerance.

Where 2 nodes coexists 4 duplicate vectors are returned.

Regards,
CVH

1) Line with command prompt in Command Line widget is missing padding/margin - letter C is glued to frame.
2) In the “Selection Filter” widget there is an “whitespace” character before colon “:” symbol on multiple places, it doesn’t seem repeat anywhere else

I use CTRL + I to pull up drawing preferences. Upon changing the dimension settings quickly and closing it normally by pressing enter it closes. But if i go to the next tab (drawing tab) then press enter to load it up quickly instead of pressing CTRL + I. Upon changing the dimensions settings and pressing enter to close it, it won’t close but instead flash and i have to keep pressing enter and eventually it closes.

I think this is a bug.

I have the Pro Version

The Application Preference setting “Graphics View - Appearance - Prevent white-on-white/black-on-black display” seems to be a global setting that doesn’t take the draw order into account. For example: With the setting enabled, it prevents creating a white hatch of a higher draw order on top of a larger, colored hatch with lower draw order (background). The preference does not seem to evaluate if the object is truly drawn as “white-on-white”.

With the preference enabled, setting the object color as “white” will set the “white” color label, but the object color actually stays black. Its hard to determine why it happens in that case.

Is it possible to let the global preference setting be applied to objects after evaluation of the draw order, and determine if the object actually is a true case of white-on-white?

My workaround: Disable the preference setting (I think it is default = on), and only enable it for ad-hoc checks of hidden and overlooked objects.

no task description

Hi Andrew

In QcadCAM when using the new Themes or User Created Themes the Layer and Block Tabs Appear to be reversed in East and West Positions.
I have discussed this with forum user PeteVick and he cannot reproduce this behavior in Qcad Pro... he suggested submitting this bug report.
I didn’t notice this until he pointed it out and don’t expect it to be high on your list of things to do..

Best Regards
John Hyslop
PS We are both using Linux Mint 20.1 (64bit) which is based on Ubuntu

Support proxy icon in title bar. This can be used to move the file to another folder using drag and drop.

QCAD Professional
Win7 WinInstaller 32bit
Version: 3.25.2.7 (3.25.2.7)
Internet: QCAD.org
Build Date: Dec 16 2020
Revision: e06f230
Qt Version: 5.13.2
Architecture: i386
Compiler: Unknown

Related forum Topic:
https://qcad.org/rsforum/viewtopic.php?t=8015

CVH

QCAD Professional
Windows 7; Win installer 32 bit
Version: 3.25.2.7 (3.25.2.7)
Internet: QCAD.org
Build Date: Dec 16 2020
Revision: e06f230
Qt Version: 5.13.2
Architecture: i386
Compiler: Unknown

Forum Topic: https://qcad.org/rsforum/viewtopic.php?t=7988
File included.
I have no intention to make 48 separate reports.

CVH

QCAD Professional
Win installer 32bit
Version: 3.25.2.4 (3.25.2.4)
Internet: QCAD.org
Build Date: Nov 18 2020
Revision: 40f4475
Qt Version: 5.13.2
Architecture: i386
Compiler: Unknown

Most likely not related to the QCAD version, rather an old sore.

File included
Topic: https://qcad.org/rsforum/viewtopic.php?f=33&t=7889#p30803
Confirmed by ‘cccplex’ Linux ‘Latest Ubuntu’

Polyline arcs with Width = 2R render as circle segments rather than circle sectors.
As Non-reversed arcs these don’t explode.
As reversed arc these explode to a single arc rather than a circle sector contour.

There is also still the faulty rendering with Z not zero
https://qcad.org/bugtracker/index.php?do=details&task_id=1923

Regards,
CVH

QCAD Professional
win installer 32bit
Version: 3.25.2.4 (3.25.2.4)
Internet: QCAD.org
Build Date: Nov 18 2020
Revision: 40f4475
Qt Version:5.13.2
Architecture: i386
Compiler: Unknown

Continued from https://qcad.org/bugtracker/index.php?do=details&task_id=1757
Splitting the old report up ...
Same file.
In the left example one can auto trim (AX):
- The lower loop.
- The top loop.
Bug?
The action on the top loop can’t be undone.

Regards
CVH

QCAD Professional
win installer 32bit
Version: 3.25.2.4 (3.25.2.4)
Internet: QCAD.org
Build Date: Nov 18 2020
Revision: 40f4475
Qt Version:5.13.2
Architecture: i386
Compiler: Unknown

Continued from https://qcad.org/bugtracker/index.php?do=details&task_id=1757
Splitting the old report up ...
Same file.
In the left example one can break out segment (D2):
- removing one of the loops with the ‘remove’ option.
- leaving both without the ‘remove’ option.
Bug?
Without the ‘remove’ option it can’t be undone.

Regards
CVH

QCAD Professional
win installer 32bit
Version: 3.25.2.4 (3.25.2.4)
Internet: QCAD.org
Build Date: Nov 18 2020
Revision: 40f4475
Qt Version:5.13.2
Architecture: i386
Compiler: Unknown

Continued from https://qcad.org/bugtracker/index.php?do=details&task_id=1757
Splitting the old report up ...
Same file.
In the left example one can cut (DI) once.
Cutting the second segment results in NO cut.

Regards
CVH

QCAD Professional
Win installer 32bit
Version: 3.25.2.4 (3.25.2.4)
Internet: QCAD.org
Build Date: Nov 18 2020
Revision: 40f4475
Qt Version: 5.13.2
Architecture: i386
Compiler: Unknown

Related topic: https://qcad.org/rsforum/viewtopic.php?f=33&t=7882
Explanation found in REntity.cpp
Listed as a Public Member Functions inherited from RObject

Regards,
CVH

Summary: Block zoomstate supersedes Model_Space after Undo of BC

QCAD Professional
Version: 3.25.2.1 (3.25.2.1)
Internet: QCAD.org
Build Date: Oct 15 2020
Revision: 2895712
Qt Version: 5.13.2
Architecture: i386
Compiler: Unknown «< (Win installer 32 bit)

With the file provided:
ZW to see the top green ‘revision cloud’ up close
QQ
Box select around the cloud
»> 2 entities in selection
BC, select origin somewhere inside the selection, accept proposed naming
»> ‘Block 1’ is created from the selection
TN
BD, select the newly created Block Reference
»> Auto zoom of ‘Block 1’ (Intended behavior)
OO
»> Returns to Model_Space with the zoomstate of the removed Block
»> Halfway down the drawing
»> The former zoomstate of the Model_Space is lost

Intended behavior is that the zoomstate of Model_Space and Blocks don’t mix
See: https://www.qcad.org/rsforum/viewtopic.php?f=31&t=7797&p=30371#p30378

This proves they do, even if the block doesn’t exists anymore.

Regards,
CVH

PS: 3th trial to enter a suggestion ...

I find myself with reasonable regularity needing to change the background color between black and white. It would be much more efficient to have a toggle button on the ribbon to make this sort of instant (kind of like the “screen based line types” button works but make it for background color and just between black and white or let people define their dark and light themes and toggle the background between the two.

When I’m drawing inside a layout block the offset tool picks up lines from blocks that are hidden.
It is not happening with all the block.

With default autosnap settings reference points included I find it very hard to snap to the baseline reference points of a dimension.
I manage to do it by setting snap to reference point only and zoom out till I see the entire dimension element.

Open QCAD and the application preferences. In the section “General: tool settings " you can resize the columns: Menu, Shortcuts etc.
The problem is that it is not intuitive. The first divider is not selectable and the second moves the first, only backwards.

How to reproduce.

1. Create a text field with some non-english words in the upper index. (see attached pic1.png)
2. Save the project and close QCAD.
3. Run QCAD and open same project.
4. Now upper/lower indexed text becomes not human readable UTF codes. (see attached pic2.png)

Tried different fonts - doesn’t help.

Linux ntb-kirill 3.10.0-1062.12.1.el7.x86_64 #1 SMP Tue Feb 4 23:02:59 UTC 2020 x86_64 x86_64 x86_64 GNU/Linux
CentOS Linux release 7.7.1908 (Core)

QCAD 3.24.2 / macOS x86_64 10.14.6

Starting with empty doc:
1. Draw any object (e.g. a rectangle) on layer 0.
2. Create new layer (”1”) with color, say yellow.
3. Select new layer
4. Select new object
5. Modify → Move / Copy
6. Click reference point
7. Click target point same as reference point
8. In “Move / Copy Options” dialog, select “Delete Original” and check “Use current layer and attributes.” Click OK.

The moved object stays on the layer it was previously on. Using “Move / Copy” as a quick way to move a selection between layers is not useful even though it appears to be the recommended way in a Google search.

Note: if “Keep Original” is selected instead, the copied object is placed on the current layer and the original object remains on its layer, as expected.

Note: if the target point is different from the reference point, the moved object is placed on the current layer as expected.

There are other workarounds too, cut / paste with “To current layer” checked seems to work, as does chaging the layer in the Properties Editor.

Openning a big file or doing any action in it, takes minutes to load.
Even dragging the scrollbar slightly left or right, takes mintues to redraw the canvas again, which makes QCAD un-usable for professional use.
Is there a way to make QCAD usable, after loading completed:
- Like suppressing edit until user says so,
- or taking snapshots of rendered canvases to another file and use it, as windows uses thumb.db ( for negative latency :P )

PS: I wish I could help a bit, but my hand are tied. I am just a java freak, I am opt to any idea of porting it to java, if she is still alive.

Tuğalsan Karabacak
EEE, MSc Mobile Comp & App, IWE

Building with the bundled version, 1.8.5, works fine.

When I try to build using the system version, which for Mageia Cauldron is 1.9.3, the build fails.

https://raw.githubusercontent.com/kekePower/mmbl/master/qcad/2019/12/11/20%3A13%3A48/log.qcad

Here is the current spec file for version 3.23.0.4.
https://svnweb.mageia.org/packages/cauldron/qcad/current/SPECS/qcad.spec?revision=1465894&view=markup

It looks like qcad is incompatible with versions above 1.8.5.

Are there plans to port to the new version?

Thanks in advance.

Yesterday I set up a Windows 64 bit install of QCAD Pro which shows the same bug I have been seeing for a while in the Community linux git repo.

Create a new drawing and then go to “Print preview”. Enable “Move paper postion”, the cursor changes to the hand icon, but the paper position is unable to be moved. Only by saving the drawing, exiting the program, restarting and reopening the drawing, will the “Move paper postion” function correctly.

Affected versions tested:

Windows 64bit installer download, as of yesterday.
Linux 64bit build from git, about 20minutes ago.

For what it’s worth, this bug is not present in this previous git checkout:

Version:
3.21.1.0 (3.21.1)
Internet:
QCAD.org
Build Date:
Jul 2 2018
Revision:

Qt Version:
5.7.0
Architecture:
x86_64
Compiler:
gcc 6.3.1

Thanks for a great product.

Win7pro 32bit Qcad 3.23.0

Andrew,
The Arc segments of Polylines with Global or any Local Width render faulty where Global Z or any of the arc’s local Z is not zero.

In the attached drawing setting the Global Z of the inner poly to 5,10,...

The setting of Z to not zero affects the radius of the projected arc and not the bulging of the poly.
It also affects the projected widths of the line inversely, but not by changing the widths properties.

Pritty rare combination! True.
Qcad is 2D native! True.

Flatten will resolve this.

It is not a purely visual rendering fault, the poly will explode as such.

Maybe similar flaw as with uneven scaling blocks with arcs.
https://qcad.org/rsforum/viewtopic.php?f=8&t=5566#p20703

Regards