Loading any larger dxf file takes very long because of this line:
istr.imbue(std::locale(”C”));

The call to std::locale is very slow. File-loading can be speed up by approximately factor 5 times on my system when replacing it. I just removed that line and instead added the following in DL_Dxf::in so it is changed only once per file:

std::locale oldLocale = std::locale::global(std::locale(”C”)); // use dot in numbers
while (readDxfGroups(fp, creationInterface)) {}
std::locale::global(oldLocale);

See also:
http://www.qcad.org/rsforum/viewtopic.php?f=33&t=3838

Splines with hundreds of control points or fit points can cause hangs when using auto snap near them.

Merely clicking on ‘Draw Order’, and usualy unintended, halts the system for several seconds up to several tens of seconds.

Intel(R) Core(TM) i7-2600K CPU @ 3.40Ghz
4,00 GB RAM installed 3,49 GB usable
32-bit Operating System
Windows 7 Professional SP1

When switching from one drawing to another, the 3D view is updated which can be slow, especially for drawings with many text entities.

See also:
https://www.qcad.org/rsforum/viewtopic.php?f=89&t=4945

Hi!

when working with files containing many not connected elements (non-poly-line elements) a double click can cause a coffee-break easily. The search for connected entities easily takes over 5 minutes on an up to date machine ([email protected]) with files that are pretty standard in size (2 or 3 MB file size ascii dxf). Adding an interruption-method would be suitable (for example by pressing escape).

attached a file with a polyline of 13000 entities and a second file with just everything duplicated. When exploding the polyline and double clicking an element it takes quite a while until all elements are selected. (this is not an extreme example, I had sometimes 10 minutes of waiting with slightly bigger files, accidently double clicking something).

also an improved algorithm for the search would help to speed up the selection. (e.g. divide the area in smaller parts, do the search there, or an option to search only in the visible view port of the document)

Martin

Hallo Andrew,

Fiddling with the drill-toolpath error, a solution seems to be to add the drill location as a ‘single point’ in the ‘source’ sublayer. I am using this as a Bug Work Around.

See also:
https://www.qcad.org/rsforum/viewtopic.php?f=33&t=6672

In files exported with the (even PG) exporter I sometimes see values like these: 334.99999999999994 or 570.0000000000001 and very often 12 digits of precision.

To safe on file size, I suggest, that QCAD rounds values on SVG export to eg. three places. When viewed in an ordinary web browser, that will result in a precision of 1 inch / 96dpi / 1000 = 0.00026mm, not? Ends should still meet.

If I understand correctly, SVG export uses svg groups to keep QCAD block entities. This in itself is very useful. There is no way to correctly keep layer information, as QCAD blocks can contain elements on different layers, while SVG only knows about groups. That is a deficiency of SVG.

To save on file size, there was some potential to get rid of ever repeating same attributes on entities, eg: style=”stroke:#000000;stroke-width:0.25;stroke-dasharray:2.16,1.08”

To optimize “By Block” is easy: just set the attribute on the SVG group, and omit on below entities styled “By Block”.

To optimize “By Layer” is a little more involved, but might be solved by using css:
- Create a class by the name of the Layer
- add the class attribute to elements styled “By Layer”

Then only individually styled entities would get a style attribute, and filesize of SVG could be halved in most cases.

This just an idea :) Having written this, I now realize, that QCAD can style weight, pattern, color each extra, so the optimazition could only be applied to entities, where everything is “By Layer” or “By Block”. But that should be the case most of the time.

add parameter to REntity::clone to allow cloning without cloning object ID.

QMdiChild widgets in an QMdiArea receive repaint and mouse events even if they are not active which slows down the user interface when having multiple MDI children open.

Long term, QMdiArea might have to be replaced with something else.

I am considering pushing the following diff:

--- a/src/core/RSPlatform.cpp
+++ b/src/core/RSPlatform.cpp
@@ -205,25 +205,6 @@ int RS::getCpuCores() {
     cores = sysinfo.dwNumberOfProcessors;
 #elif defined(Q_OS_UNIX)
     cores = sysconf( _SC_NPROCESSORS_ONLN );
-#elif defined (Q_OS_MAC) || defined (Q_OS_BSD4) || [...]
-    nt mib[4];
-    size_t len;
-
-    /* set the mib for hw.ncpu */
-    mib[0] = CTL_HW;
-    mib[1] = HW_AVAILCPU; // alternatively, try HW_NCPU;
-
-    /* get the number of CPUs from the system */
-    sysctl(mib, 2, &cores, &len, NULL, 0);
-
-    if (cores < 1) {
-        mib[1] = HW_NCPU;
-        sysctl( mib, 2, &cores, &len, NULL, 0 );
-    }
-#elif defined(Q_OS_HPUX)
-    cores = mpctl(MPC_GETNUMSPUS, NULL, NULL);
-#elif defined(Q_OS_IRIX)
-    cores = sysconf( _SC_NPROC_ONLN );
 #endif

(truncated list to manage size)

I see that most of this was put in there as a kind of a note-to-self originally, however _SC_NPROCESSORS_ONLN is implemented by every realistically possible targets anyway, so for all practical purposes, this is dead weight.

A couple of questions remain, though:

• #ifdef Q_OS_WIN32 should probably be Q_OS_WIN, so as to include win64 as well.
• Assuming the real purpose for this function is to determine how many parts a particular task is to be split into in the future, if and when QCAD gets the need to split expensive tasks among multiple CPUS, “core” probably shouldn’t be static - CPUs are hot-pluggable today, and while one probably doesn’t do that on a laptop, it’s not unfeasible in, say, a VDI environment.

Thoughts?

Both of these files (scripts/simple.js and scripts/library.js) contain a function named sleep() that duplicate each other. There is probably a need to delete one of them.

Permalinks to this definitions on Github:

• simple.js
• library.js

Disable

Adjust Version

Application

Deployment

Don't forget

Check List

Ports

I only noticed today (perhaps they were available earlier?) that there are 64 bit binaries for windows; I suppose these are true 64 bit.

Yet, the installer prompts me to put them into “Programs (x86)”, where all the 32bit software lives, I guess - this seems to be a pure convention though, the application runs just fine.

Maybe the reason is, that the 32bit version was installed before; Maybe it should change? Maybe it should stay, so that there are not two?

https://qcad.org/rsforum/viewtopic.php?f=30&t=6547&p=29307#p29307

FlexPainter is updated.

Should not load while PROTOOLS are absent.

Regards,
CVH

 FS#1604  was issued because of non-intuitive hierarchy of the preferences for the Explode function.
The task comment recieved explains the funcionallity in details not found elsewhere.

Indentation of the phrase ‘Spline Segments:’ suggest hierarchy.
Graying out the drop down box for the number of spline segments establish hierarchy.

QCADpro 3.25.2.1 / win7pro 32 bit / 4GB / 1e generation i7 2.6Ghz / HDD GeForce GT 520

The method PE casts the clipboard on individual hatch segments.
Hatches should be considered as a single complex entity not allowing certain actions.

Discovered while:
https://qcad.org/rsforum/viewtopic.php?f=33&p=30008#p30008

Not that it matters: In doing so the scaling is off too.

Regards,
CVH

Andrew,
unrelated to any OS or version.

Circles do explode to polylines with two arc segments.
Discussed earlier and very correct because a full circle bulge would be a mathematical limit situation.

When there is no option to ‘Undo’ anymore or when the file was saved & loaded there is no easy way back.

One of the reasons why I isolate circle entities if I might going to use XP as in the file here:
https://qcad.org/bugtracker/index.php?do=details&task_id=2155
Even knowing it, I still make more mistakes as I would like.

It would be nice to be able to exclude circle entities from the GUI explosion action XP.
Can be simply implemented in the top of Explode.explodeSelection with an extra preference in AppPrefs, default =true.

On the other hand, the explosions of circles used internally in various methods should remain.

Regards,
CVH

Hi,

Using the Copy and Move command the [ ] Use current layer and attributes’ is confusing.

Perhaps [ ] change copy /moved set to current layers and attributes. is more suitable.

After each update, I add a dot before the ~user/opt folder (and in the program launcher) created by QCAD to unclutter my file manager’s appearance. Would be easy, I think, to make the installer put the program directly in ~user/.opt, and nobody would miss that folder while browsing through daily used files.

And what about installing the program as root, so all users of a computer could use it without personal installs, and only personal setting files created by users would be kept in ~user/.opt/qcam.

Jean-Claude

For all the Linux VI editor geeks, can you disable the <ESC> function in the text edit mode. <ESC> maps to <cancel>. If anything it should map to <ok>. Which creates no harm as it can be later undone. I have accidentally lost edits countless times due to an unbreakable habit.

As it says on the tin - please add script generator for Qt 5.6.2.

The startup script still forces a QT-style; Getting bored of that one, qcad runs fine with the system default style - probably that was a qt4 thing, that can be removed, so it blends better with the DE?

18c18 /opt/qcad/qcad
< LD_LIBRARY_PATH=”$DIR” “$binary” -style plastique “[email protected]” —

LD_LIBRARY_PATH=”$DIR” “$binary” “[email protected]”

... for the Option bar.

Since the launcher batch no longer requests “-style plastique” QT wants to apply the native style of the desktop environment. This fails on a headless system (or where the current user has no graphical shell to connect to) with the message “Gtk-WARNING **: cannot open display: “. Fortunately, QT provides a workaround in addtion to the new “platform=offscreen” flag.

In the tools that set -no-gui, beforehand export these variables into the environment

export QT_STYLE_OVERRIDE=""
export QT_QPA_PLATFORM=offscreen
export DISPLAY=""

PS: QT-offscreen will not find fonts; unfortunately the workaround only allows for a single directory to search them (I need helvetica):

export QT_QPA_FONTDIR="/usr/local/share/fonts/type1"

Maybe that should get reported upstream, so that like LD_LIBRARY_PATH several directories could be specified.

Is it possible to create “OVERKILL” command like in AUTOCAD

Please add script bindings for Qt 5.9.4, thank you.

Without any selection.
Entities on a locked layer can be selected after choosing:

Asumed wrong
- Modify DI
- Modify B2
- Modify D3
- Modify MT

Questionable
- Modify RM
- Modify CH
- Modify RN Less severe
- Modify OF
- Modify OH
- Modify SS

Regards