Re: [Inkscape-devel] [Bug 307195] [NEW] crash with two inkscape instances when changing layer
by the Adib
Bulia,
I attached the patch to the bug tracker.
The problem is easy to reproduce as described in the tracker.
pls check. Adib.
---
On Thu, Sep 10, 2009 at 12:36 AM, Launchpad Bug
Tracker<307195@...1882...> wrote:
> You have been subscribed to a public bug:
>
> Steps to reproduce:
>
> * open first file
> * open second file
> * in first window show layers dialog
> * in second window select (click) some object on canvas
> * in first window change layer
> -> crash
>
> Linux, SVN Revision: 20351
>
> Program received signal SIGSEGV, Segmentation fault.
> [Switching to Thread -1247754000 (LWP 10472)]
> 0x081b5e33 in SPDesktop::currentRoot () at io/base64stream.cpp:74
> 74 */
> Current language: auto; currently c++
> (gdb) backtrace
> #0 0x081b5e33 in SPDesktop::currentRoot () at io/base64stream.cpp:74
> #1 0x083eb175 in Inkscape::UI::Dialogs::LayersPanel::_pushTreeSelectionToCurrent () at io/base64stream.cpp:74
> #2 0x083f3ff3 in sigc::bound_mem_functor0<void, Inkscape::UI::Dialogs::LayersPanel>::operator() () at io/base64stream.cpp:74
> #3 0x083f400a in sigc::adaptor_functor<sigc::bound_mem_functor0<void, Inkscape::UI::Dialogs::LayersPanel> >::operator() () at io/base64stream.cpp:74
> #4 0x083f4026 in sigc::internal::slot_call0<sigc::bound_mem_functor0<void, Inkscape::UI::Dialogs::LayersPanel>, void>::call_it () at io/base64stream.cpp:74
> #5 0xb7b83526 in Glib::SignalProxyNormal::slot0_void_callback () from /usr/lib/libglibmm-2.4.so.1
> #6 0xb76f2207 in g_cclosure_marshal_VOID__VOID () from /usr/lib/libgobject-2.0.so.0
> #7 0x00000000 in ?? ()
> (gdb)
>
> More precisely, the access on _layer_hierarchy seems to produce the
> crash.
>
> SPObject *SPDesktop::currentRoot() const
> {
> return _layer_hierarchy ? // <- crashes here
> _layer_hierarchy->top() : NULL;
> }
>
> ** Affects: inkscape
> Importance: High
> Assignee: theAdib (theadib)
> Status: In Progress
>
>
> ** Tags: crash layers regression
> --
> crash with two inkscape instances when changing layer
> https://bugs.launchpad.net/bugs/307195
> You received this bug notification because you are a bug assignee.
>
13 years, 8 months
Remapping CNTL key.
by Karl Edler
Hello,
I use an alpha-grip (http://www.alphagrips.com/ ) instead of a normal
keyboard and this makes it very hard to perform a CTL-click operation. I
tried remapping the CTL button using xkb (KDE's keymapping tool) but the
settings did not translate to inkscape because (I think) it uses its own
keymapping file "default.xml". How might I remap the keyboard under
inkscape so that pressing the 'g' key gives the same function as
pressing CTL used to? Then I could 'g'-click rather than "CTL-click". Is
this possible or is programming necessary before this is possible?
I could not find this information on the wiki. If it is there please
just point the way.
Thanks.
13 years, 8 months
PNG Compression
by Joshua A. Andler
Hey there,
I just wanted to point out that the "crunching" done on our pngs
produces files that have forced alpha channels. Obviously if the image
needs transparency, no harm there. However, for things that should have
no opacity (such as screenshots), it can end up bloating them more than
they need to be.
Also, I found something that compresses better than pngcrush (however,
it is MUCH slower). It's called "pngout", it's cross-platform, it's
free, but it is not open source. I'm going to be re-compressing all of
our screenshots with it because the results are much better (imho).
However, I want to make sure that we get someone with IE to check that
they view fine there before proceeding. They do load okay in Opera 10,
Firefox 3.5, and everything else that I checked.
A comparison of the 0.47 "spiro" screenshot was as follows:
355726 - Version currently on server
349063 - Saved from GIMP with the unnecessary alpha channel removed
337597 - Compressed w/ pngcrush (brute option)
327638 - Compressed w/ pngout (RGB color option)
28k is 28k... better for serving and receiving. In the end, doing this
to all of our screenshots will probably add up to quite a bit.
If anyone is interested in using pngout, it can be obtained here:
http://advsys.net/ken/utils.htm
Note: Someone else ported it to Linux, OSX, and BSD... so you'll need to
follow a link on that page to the next place.
Cheers,
Josh
13 years, 8 months
Fixed point arithmetic in libnr causes pixel offset at large zoom
by Diederik van Lierop
Hi all,
As reported in https://bugs.launchpad.net/inkscape/+bug/168384, bitmap
pixels appear slightly offset when nearing the maximum zoom.
I found out that this is caused by the fixed point calculations in
libnr, for example in nr_R8G8B8A8_P_R8G8B8A8_P_R8G8B8A8_N_TRANSFORM_0().
This is easily fixed by increasing the fixed point precision from 12 to
18 bits on line 42 in libr/nr-compose-transform.cpp, but if we do that
then our headroom will become quite small. Can we switch to long long
instead of just long? Or should we go to floating point math, and accept
the speed penalty (if any)? Or should we leave things as they are?
Someone must have though about this before ;-)
Diederik
13 years, 8 months
Re: [Inkscape-devel] Inkscape 0.47pre2
by Chris Morgan
Forgot to reply to all again... anyway, new news is
http://inkscape.chrismorgan.info/win32-builds/Inkscape-0.47pre2-2.exe with
the updated devlibs. I'd especially like this tested with graphics tablets
to see if the updated GTK version (not the newest, but newer) fixes that
problem with them not working as extended input devices.
On Tue, Sep 1, 2009 at 9:49 PM, Chris Morgan <chris.morganiser@...400...>wrote:
> On Tue, Sep 1, 2009 at 8:02 AM, Alvin Penner <penner@...1856...> wrote:
>
>> Alexandre Prokoudine wrote:
>> >
>> > OK, so can we use this build?
>> >
>> I've been using the build :
>> inkscape.chrismorgan.info/win32-builds/Inkscape-0.47pre2-1.exe.
>> and it is working very well for me, thanks Chris!
>>
> Good :-)
>
> The only 'issue' I've seen is that it apparently does not include the
>> Python
>> packages 'PIL' and 'uniconvertor', which I think would normally be present
>> in a Windows installation.
>>
> I happened to notice that today before you emailed, looking at it for some
> PortableApps.com stuff (CommonFiles stuff - a "feasibility study" for a
> unified Python distribution except for extra libs, for Python apps and apps
> that use Python, e.g. Blender, Inkscape (both contain), Task Coach, BPBible,
> others (are written in)). I'd seen PIL in 0.46, and noticed it wasn't
> there. It has since occurred to me that I have not updated my devlibs
> bundle (the download contains all the SVN metadata for the repo) as
> recommended in the Win32Port wiki page, so I'm going to do that now and
> rebuild it all. I think that that should fix it, and also it'll just make
> sure everything is up to date. The devlibs bundle I downloaded was at r14,
> so I'm updating it to the latest, r20.
>
> Sorry about that guys. I plan now on being able to upload
> Inkscape-0.47pre2-2.exe tomorrow (probably in about 20 hours time).
>
> --
>> View this message in context:
>> http://www.nabble.com/Inkscape-0.47pre2-tp25110580p25231102.html
>> Sent from the Inkscape - Dev mailing list archive at Nabble.com.
>>
>>
>>
>> ------------------------------------------------------------------------------
>> Let Crystal Reports handle the reporting - Free Crystal Reports 2008
>> 30-Day
>> trial. Simplify your report design, integration and deployment - and focus
>> on
>> what you do best, core application coding. Discover what's new with
>> Crystal Reports now. http://p.sf.net/sfu/bobj-july
>> _______________________________________________
>> Inkscape-devel mailing list
>> Inkscape-devel(a)lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/inkscape-devel
>>
>
>
>
> --
> God bless,
>
> Chris Morgan <chris.morganiser@...400...>
>
>
> I'm good at making two things: mistakes and enemies.
>
-- Chris Morgan <chris.morganiser@...400...>
I'm good at making two things: mistakes and enemies.
13 years, 9 months
Little bug in i18n
by Lucas Vieites
Hi,
I'm testing and debugging my translation on the inkscape-0.47pre2 version
and I found the following:
When I apply a filter to an object, say "Wrinkled varnish", the menu
option appears correctly translated as "Barniz arrugado", but when I select
the object, on the status bar I see the description of the object, for
example:* "Elipse; **filtrado (Wrinkled varnish)* en capa* Capa 1*.
Pulse......". As you see, the name of the filter does not appear translated
here.
I think it's this segment in the po file: the second "%s" is not getting
the correct string.
============================
#: ../src/sp-item.cpp:1065
#, c-format
msgid "%s; <i>filtered (%s)</i>"
msgstr "%s; <i>filtrado (%s)</i>"
============================
Should I open a bug report?
Thanks,
--
Lucas Vieites
13 years, 9 months
Windows installer change request
by Jon
Just downloaded Inkscape-0.47pre2-1.exe and am trying it out.
Please consider modifying the installer to provide an option to allow one to
install in an existing directory that has files and subdirs. The use case
for this request is that I customize the Inkscape theme on my systems and
therefore have content in etc, lib, and share subdirs.
I think a "safe" solution is to provide an "Overwrite!" button in the dialog
that pops up when you try to install to a non-empty dir. Keep the "Retry"
and "Cancel" buttons and default to not allowing one to install to a
non-empty dir, just provide us of those who have existing customizations the
option of continuing the installation without having to waste time copying
the customizations to a temp directory and then copying them back.
Thanks, and THANKS for all the coolness that is Inkscape!
Jon
13 years, 9 months
Re: [Inkscape-devel] Win32 libs update request - was Re: Inkscape0.47pre2
by Preben Soeberg
> -----Original Message-----
> From: Preben Soeberg [mailto:prsodk@...400...]
> Sent: 04 September, 2009 20:32
> To: 'J.B.C.Engelen@...1578...'
> Subject: RE: [Inkscape-devel] Win32 libs update request - was Re:
> Inkscape0.47pre2
>
> ...
> > >
> > > It has been much improved, but bug
> > > https://bugs.launchpad.net/bugs/284546 is still there, though
> > > it does not appear quite as easy as before.
> > >
> > > Couldn't the THROW_CONTINUITYERROR at line 355 in
> > > src\2geom\path.cpp be commented out, as suggested long time
> > > ago. That would help a lot.
> >
> > We should definitely not remove THROW_CONTINUITYERROR ! It means there
> > is a real problem that Inkscape/2geom should not ignore.
> > I think what causes the problem in this case is adding a line segment
> > with 'invalid' points (like infinity). Perhaps adding a simple check for
> > that in the callygr tool will fix this issue.
> >
> > Ciao,
> > Johan
>
> Yes, it's a dirty workaround I suggested, but it works.
> I commented it out and have been using it like that for 8 months now with
> only minor problems. These minor problems consist in, that when I
> calligraphic-trace and move the pen too slow, the trace is aborted, but
> without the crash.
>
> I don't have time to try to look for a way to fix it (for the next week or
> so). I hope somebody else is trying.
>
> Preben
13 years, 9 months
