Cursor placement regression (painting?)

Post by Scott Kostyshak
With Qt4 and Qt5, if I open the attached document I cannot place the
cursor to the right of the note. If I put the cursor to the left of
the BibTeX inset and press left, it goes to the left of the note, if I
press left again it enters the note from the right.
The cursor placement works as expected for me at 1fefbf5b.

This should be fixed now.

JMarc

Scott Kostyshak

2014-07-29 01:33:34 UTC

Post by Jean-Marc Lasgouttes

On Mon, Jul 28, 2014 at 4:20 PM, Jean-Marc Lasgouttes

This should be fixed now.

Works well.

Scott

Enrico Forestieri

2014-08-01 11:49:50 UTC

Post by Jean-Marc Lasgouttes

Post by Scott Kostyshak
On Mon, Jul 28, 2014 at 4:20 PM, Jean-Marc Lasgouttes

This should be fixed now.

Works well.

If I put the cursor at the end of the first line (after the note inset) and
then close the document, LyX segfaults. Running under gdb it only crashes
the second time, which is weird. Here is the backtrace:

Program received signal SIGSEGV, Segmentation fault.
0x081654e4 in lyx::Cursor::selection (this=0x158) at ../../src/Cursor.h:166
166 bool selection() const { return selection_; }
(gdb) bt
#0 0x081654e4 in lyx::Cursor::selection (this=0x158) at ../../src/Cursor.h:166
#1 0x08525650 in lyx::frontend::GuiView::dispatch (this=0x9066718, cmd=...,
dr=...) at ../../../../src/frontends/qt4/GuiView.cpp:3840
#2 0x084de973 in lyx::frontend::GuiApplication::dispatch (this=0x8c1e948,
cmd=..., dr=...) at ../../../../src/frontends/qt4/GuiApplication.cpp:1973
#3 0x084daf4e in lyx::frontend::GuiApplication::dispatch (this=0x8c1e948,
cmd=...) at ../../../../src/frontends/qt4/GuiApplication.cpp:1332
#4 0x081eaed7 in lyx::dispatch (action=...) at ../../src/LyX.cpp:1303
#5 0x0857abb6 in lyx::frontend::Action::action (this=0x92993f8)
at ../../../../src/frontends/qt4/Action.cpp:65
#6 0x0857ac20 in lyx::frontend::Action::qt_static_metacall (_o=0x92993f8,
_c=QMetaObject::InvokeMetaMethod, _id=1, _a=0xbfffe3d8)
at ./moc_Action.cpp:77
#7 0xb70d86ed in QMetaObject::activate(QObject*, int, int, void**) ()
from /usr/local/qt/5.3.1/lib/libQt5Core.so.5
#8 0xb70d90db in QMetaObject::activate(QObject*, QMetaObject const*, int, void**) () from /usr/local/qt/5.3.1/lib/libQt5Core.so.5
#9 0xb790b15d in QAction::triggered(bool) ()
from /usr/local/qt/5.3.1/lib/libQt5Widgets.so.5
#10 0xb790d2f9 in QAction::activate(QAction::ActionEvent) ()
from /usr/local/qt/5.3.1/lib/libQt5Widgets.so.5
#11 0xb7ab67d4 in ?? () from /usr/local/qt/5.3.1/lib/libQt5Widgets.so.5
#12 0xb7abe94f in ?? () from /usr/local/qt/5.3.1/lib/libQt5Widgets.so.5

--
Enrico

Enrico Forestieri

2014-08-01 14:36:45 UTC

Post by Enrico Forestieri
If I put the cursor at the end of the first line (after the note inset) and
then close the document, LyX segfaults. Running under gdb it only crashes

It seems that this issue occurs only with Qt5. When compiling with Qt4 I
don't experience any crash.

--
Enrico

Enrico Forestieri

2014-08-01 15:41:28 UTC

Post by Enrico Forestieri
If I put the cursor at the end of the first line (after the note inset) and
then close the document, LyX segfaults. Running under gdb it only crashes

It seems that this issue occurs only with Qt5. When compiling with Qt4 I
don't experience any crash.

Anyway, there is something fishy going on here. Launching lyx with no document
open and selecting "Help->About LyX" I get a segfault with either Qt4 or Qt5.
This does not occur at revision 1fefbf5b.

--
Enrico

Scott Kostyshak

2014-08-01 16:10:26 UTC

Post by Enrico Forestieri
If I put the cursor at the end of the first line (after the note inset) and
then close the document, LyX segfaults. Running under gdb it only crashes

It seems that this issue occurs only with Qt5. When compiling with Qt4 I
don't experience any crash.

Anyway, there is something fishy going on here. Launching lyx with no document
open and selecting "Help->About LyX" I get a segfault with either Qt4 or Qt5.
This does not occur at revision 1fefbf5b.

I also get a crash with Help > About on Qt 5. I haven't tested on Qt 4.

Scott

Enrico Forestieri

2014-08-03 18:56:03 UTC

Post by Enrico Forestieri
If I put the cursor at the end of the first line (after the note inset) and
then close the document, LyX segfaults. Running under gdb it only crashes

It seems that this issue occurs only with Qt5. When compiling with Qt4 I
don't experience any crash.

Anyway, there is something fishy going on here. Launching lyx with no document
open and selecting "Help->About LyX" I get a segfault with either Qt4 or Qt5.
This does not occur at revision 1fefbf5b.

I also get a crash with Help > About on Qt 5. I haven't tested on Qt 4.

I took the time to bisect the thing and the culprit turns out to be
revision fb05011a

commit fb05011ad0db2e2c5e557d91f398970d75cf8d15
Author: Scott Kostyshak <***@lyx.org>
Date: Wed Jun 25 16:50:31 2014 -0400

Despite the date, this was committed a few days ago.

--
Enrico

Scott Kostyshak

2014-08-03 22:53:54 UTC

Post by Enrico Forestieri
If I put the cursor at the end of the first line (after the note inset) and
then close the document, LyX segfaults. Running under gdb it only crashes

It seems that this issue occurs only with Qt5. When compiling with Qt4 I
don't experience any crash.

Anyway, there is something fishy going on here. Launching lyx with no document
open and selecting "Help->About LyX" I get a segfault with either Qt4 or Qt5.
This does not occur at revision 1fefbf5b.

I also get a crash with Help > About on Qt 5. I haven't tested on Qt 4.

I took the time to bisect the thing and the culprit turns out to be
revision fb05011a

Thanks for doing the bisect Enrico. I confirm and don't know why it
happens. I will revert.

Post by Enrico Forestieri
commit fb05011ad0db2e2c5e557d91f398970d75cf8d15
Date: Wed Jun 25 16:50:31 2014 -0400
Despite the date, this was committed a few days ago.

It was suggested to me to keep the author date from when the patch was
first made public and I liked the suggestion. Let me know if you find this
confusing and we can discuss it. Note also that there is a commit date and
you can choose to display that instead of or in addition to the author date.

Scott

Scott Kostyshak

2014-08-03 23:23:43 UTC

Thanks for doing the bisect Enrico. I confirm and don't know why it happens.
I will revert.

Reverted at 453ce611.

Scott

Enrico Forestieri

2014-08-03 23:42:59 UTC

Post by Enrico Forestieri
I took the time to bisect the thing and the culprit turns out to be
revision fb05011a

Thanks for doing the bisect Enrico. I confirm and don't know why it
happens. I will revert.

I think you will also have to close a flurry of bugs opened by Uwe and
caused by that commit ;)

Post by Enrico Forestieri
commit fb05011ad0db2e2c5e557d91f398970d75cf8d15
Date: Wed Jun 25 16:50:31 2014 -0400
Despite the date, this was committed a few days ago.

Uh, I really don't care but I think that the commit date would be
more useful. My comment above was only trying to say that it was
something recent.

--
Enrico

Scott Kostyshak

2014-08-04 02:53:27 UTC

Post by Enrico Forestieri
I took the time to bisect the thing and the culprit turns out to be
revision fb05011a

Thanks for doing the bisect Enrico. I confirm and don't know why it
happens. I will revert.

I think you will also have to close a flurry of bugs opened by Uwe and
caused by that commit ;)

It seems Uwe has closed them.