I've created one ellipse and I tried to divide it into two pieces with a line.
The first (TOP) point works the second causes a seg fault.
I am running on Fedora 20 with version 2.0.6
(gdb) bt
0 0x00000000004c7482 in ()
1 0x00000000004cbe93 in ()
2 0x000000000055f108 in ()
3 0x00000000005caf08 in ()
4 0x00000000005cb665 in ()
5 0x0000000000535669 in ()
6 0x000000000053b5a7 in ()
7 0x0000000000601b7c in ()
8 0x00007f89b09ab118 in QWidget::event(QEvent) (this=0x1617c70, event=0x7fffef599f60)
at kernel/qwidget.cpp:8775
9 0x00007f89b0957efc in QApplicationPrivate::notify_helper(QObject, QEvent) (this=this@entry=0x1154dc0, receiver=receiver@entry=0x1617c70, e=e@entry=0x7fffef599f60) at kernel/qapplication.cpp:4565
10 0x00007f89b095eabf in QApplication::notify(QObject, QEvent*) (this=<optimized out="">, receiver=0x1617c70, e=0x7fffef599f60) at kernel/qapplication.cpp:4108</optimized>
11 0x00007f89b042722d in QCoreApplication::notifyInternal(QObject, QEvent) (this=
0x7fffef59a940, receiver=receiver@entry=0x1617c70, event=event@entry=0x7fffef599f60)
at kernel/qcoreapplication.cpp:953
at ../../src/corelib/kernel/qcoreapplication.h:231
13 0x00007f89b095e11f in QApplicationPrivate::sendMouseEvent(QWidget, QMouseEvent, QWidget, QWidget, QWidget**, QPointer<qwidget>&, bool) (receiver=receiver@entry=0x1617c70, event=event@entry=</qwidget>
0x7fffef599f60, alienWidget=alienWidget@entry=0x1617c70, nativeWidget=nativeWidget@entry=0x131fb00, buttonDown=buttonDown@entry=0x7f89b147a478 <qt_button_down>, lastMouseReceiver=..., spontaneous=true)
at kernel/qapplication.cpp:3173
14 0x00007f89b09d4562 in QETWidget::translateMouseEvent(_XEvent const*) (this=this@entry=0x131fb00, event=event@entry=0x7fffef59a2c0) at kernel/qapplication_x11.cpp:4540
15 0x00007f89b09d2f5c in QApplication::x11ProcessEvent(_XEvent*) (this=0x7fffef59a940, event=event@entry=0x7fffef59a2c0) at kernel/qapplication_x11.cpp:3663
16 0x00007f89b09faff4 in x11EventSourceDispatch(GSource*, GSourceFunc, gpointer) (s=0x1155970, callback=0x0, user_data=0x0) at kernel/qguieventdispatcher_glib.cpp:148
17 0x00007f89ae2b17fb in g_main_context_dispatch () at /lib64/libglib-2.0.so.0
18 0x00007f89ae2b1b98 in g_main_context_iterate.isra () at /lib64/libglib-2.0.so.0
19 0x00007f89ae2b1c4c in g_main_context_iteration () at /lib64/libglib-2.0.so.0
20 0x00007f89b045655e in QEventDispatcherGlib::processEvents(QFlags<qeventloop::processeventsflag>) (this=</qeventloop::processeventsflag>
0x11551c0, flags=...) at kernel/qeventdispatcher_glib.cpp:450
21 0x00007f89b09fb176 in QGuiEventDispatcherGlib::processEvents(QFlags<qeventloop::processeventsflag>) (this=<optimized out="">, flags=...) at kernel/qguieventdispatcher_glib.cpp:207</optimized></qeventloop::processeventsflag>
22 0x00007f89b0425d81 in QEventLoop::processEvents(QFlags<qeventloop::processeventsflag>) (this=this@entry=0x7fffef59a6a0, flags=...) at kernel/qeventloop.cpp:149</qeventloop::processeventsflag>
23 0x00007f89b04260e5 in QEventLoop::exec(QFlags<qeventloop::processeventsflag>) (this=this@entry=0x7fffef59a6a0, flags=...) at kernel/qeventloop.cpp:204</qeventloop::processeventsflag>
24 0x00007f89b042b7c9 in QCoreApplication::exec() () at kernel/qcoreapplication.cpp:1225
25 0x00000000004316e7 in ()
26 0x00007f89af257fe0 in __libc_start_main (main=
0x42fbe0, argc=2, argv=0x7fffef59ab38, init=<optimized out>, fini=<optimized out>, rtld_fini=<optimized out>, stack_end=0x7fffef59ab28) at libc-start.c:289
27 0x0000000000435247 in ()
Tried with LC 2.0 HEAD ( https://github.com/LibreCAD/LibreCAD/commits/2.0 ), it doesn't crash here.
Can you test it with the latest 2.0?
Rpm files for latest 2.0 brach code (2.0.8+git)
http://download.opensuse.org/repositories/home:/Rallaz/Fedora_22/
or
http://download.opensuse.org/repositories/home:/Rallaz/Fedora_21/
required packages are:
libdxfrw0-0.6.3-1.1 & librecad-2.0.8+git20151019.dcd167d-1.1
I've downloaded the latest 2.0.8+git from the above packages.
I cannot perform the cut because it doesn't detect/snap on the (lower) intersection of the line and ellipse.
Some other (minor) things
(not for this thread...)
I can reproduce the missing intersection issue, and I will fix this one.
weird enough, I can still trim the ellipse, even intersection failed.
pushed one fix for intersections in master and 2.0
commit in master: 3d1ee00
I'm curious why this is not marked a closed - fixed?
How is that supposed to happen?