F9 + KDE 4.0 (and 4.1) + Inkscape (missing color picker widgets)
Hello,
My apologies if I'm asking in the wrong place (this seemed like it might be a bug, so I thought I'd try here).
So I installed inkscape-0.46-2.fc9.x86_64, and it's exhibiting unexpected behaviour in the color-picking widgets (in the Fill + Stroke dialog). Basically, they're not displaying any color:
http://imagebin.org/24367 (RGB) http://imagebin.org/24368 (HSL) http://imagebin.org/24369 (CMYK)
The 'Wheel' color picker, however, seems just fine. Is there a specific package dependency that I need to reinstall to get these color pickers to work (this is in Fedora 9).
I also tried compiling from http://inkscape.modevia.com/svn-snap/inkscape-19664.tar.bz2 (build 19666 didn't seem to compile), but got the same result with the color pickers. Can someone help point me in the right direction to solving this?
Thanks.
--Mike
On Aug 15, 2008, at 9:58 PM, Michael Park wrote:
So I installed inkscape-0.46-2.fc9.x86_64, and it's exhibiting unexpected behaviour in the color-picking widgets (in the Fill + Stroke dialog). Basically, they're not displaying any color:
http://imagebin.org/24367 (RGB) http://imagebin.org/24368 (HSL) http://imagebin.org/24369 (CMYK)
The 'Wheel' color picker, however, seems just fine. Is there a specific package dependency that I need to reinstall to get these color pickers to work (this is in Fedora 9).
I also tried compiling from http://inkscape.modevia.com/svn-snap/inkscape-19664.tar.bz2 (build 19666 didn't seem to compile), but got the same result with the color pickers. Can someone help point me in the right direction to solving this?
That is quite interesting.
Those first three you are having problems with all use a custom slider widget that was inherited from pre-Inkscape code. The "wheel" picker is also custom code, but was written differently (by me) just after Inkscape branched.
That you see the same results with both the prebuilt version and the version you compile shows it is probably something fundamental about the approach being taken.
Can you run from the console and look for any error messages?
And if you want to dive in a little further, the code involved there is pretty self-contained and is not hard to start with.
Hi Jon,
Can you run from the console and look for any error messages?
Sure, so I tried that, and there was no asserts or any other debugging output related to the color pickers. (I drew a rectangle, then changed its colors using every color picker, using every slider, and witnessed that they did seem to affect the color in the intended way). Still no output though.. I didn't see any 'verbose' command-line option, is there one that's not published in the '-?' list?
And if you want to dive in a little further, the code involved there is pretty self-contained and is not hard to start with.
Sure, I wouldn't mind giving it a shot, although please go easy on me- my C/C++ is pretty rusty. :) Not to mention that I'm also pretty green when it comes to desktop appdev (I'm more of a webdev).
If there's any other info that I can provide, please let me know. Thanks.
--Mike
Michael Park wrote:
I also tried compiling from http://inkscape.modevia.com/svn-snap/inkscape-19664.tar.bz2 (build 19666 didn't seem to compile)
What are the errors you're seeing? Rev. '65 and '66 are mine, but they work fine for me. I'm wondering if I somehow messed up when updating to the latest revision of 2geom...
Regards,
Diederik
Hi Diederik,
What are the errors you're seeing? Rev. '65 and '66 are mine, but they work fine for me. I'm wondering if I somehow messed up when updating to the latest revision of 2geom...
Yes, it seems like it's looking for a different version of the C++ stdlib. Here's the compile error I got:
...<snip>...
2geom/geom.cpp: In function 'void Geom::eliminate_duplicates_p(std::vector<Geom::Point, std::allocatorGeom::Point >&)': 2geom/geom.cpp:243: error: 'sort' is not a member of 'std' make[2]: *** [2geom/geom.o] Error 1
...<snip>...
...and here's my version of libstd:
Installed Packages libstdc++.i386 4.3.0-8 installed libstdc++.x86_64 4.3.0-8 installed libstdc++-devel.x86_64 4.3.0-8 installed
Let me know if you want more (or the rest of the) make output, or anything else.
Thanks.
--Mike
That would come from std::sort not being defined. I would suggest adding
#include <algorithm>
to the top of that file. Possibly it is getting picked up by other includes on other people's machines, but not on yours. Including it explicitly should help. We get the same problem sometimes with <string.h> being included "automatically" on some people's systems, but not on others.
bob
Michael Park wrote:
...<snip>...
2geom/geom.cpp: In function 'void Geom::eliminate_duplicates_p(std::vector<Geom::Point, std::allocatorGeom::Point >&)': 2geom/geom.cpp:243: error: 'sort' is not a member of 'std' make[2]: *** [2geom/geom.o] Error 1
...<snip>...
Bob Jamison schrieb:
That would come from std::sort not being defined. I would suggest adding
#include <algorithm>
to the top of that file. Possibly it is getting picked up by other includes on other people's machines, but not on yours. Including it explicitly should help. We get the same problem sometimes with <string.h> being included "automatically" on some people's systems, but not on others.
Yep, that missing include comes from a change I made and it compiles just fine for me. Thanks for the hint that this is a potential problem on some machines. I now added it to geom.cpp (rev. 19672). I don't need to put this into #ifdef's as with config.h, though, do I?
Max
On Sun, Aug 17, 2008 at 12:17 PM, Maximilian Albert <Anhalter42@...173...> wrote:
Bob Jamison schrieb:
That would come from std::sort not being defined. I would suggest adding
#include <algorithm>
to the top of that file. Possibly it is getting picked up by other includes on other people's machines, but not on yours. Including it explicitly should help. We get the same problem sometimes with <string.h> being included "automatically" on some people's systems, but not on others.
Yep, that missing include comes from a change I made and it compiles just fine for me. Thanks for the hint that this is a potential problem on some machines. I now added it to geom.cpp (rev. 19672). I don't need to put this into #ifdef's as with config.h, though, do I?
Just wanted to follow up, Bob's suggestion did fix the compilation error for me as well.
Max
No, I don't think so. I didn't have that problem before, and added that line to the file, and it compiles ok.
bob
Yep, that missing include comes from a change I made and it compiles just fine for me. Thanks for the hint that this is a potential problem on some machines. I now added it to geom.cpp (rev. 19672). I don't need to put this into #ifdef's as with config.h, though, do I?
Max
Michael,
This looks to me to be a GTK-QT Theme Engine issue to me. I'm guessing that you have gtk-qt-engine-kde4 (debian/ubuntu package name) installed and Use My KDE style in Gtk apps selected. Do into the system settings->appearance->GTK styles and Fonts pick the use another style option pick a regular GTK theme will make them look normal, ugly but normal.
Again this looks like a GTK-QT Theme Engine bug.
Joshua L. Blocher verbalshadow
participants (6)
-
Bob Jamison -
Diederik van Lierop -
Jon A. Cruz -
Joshua L. Blocher -
Maximilian Albert -
Michael Park