New upstream release qtiplot 0.9.7.8 (sync with debian and fix for debian's FTBFS)

[Alessandro Ghersi]

Sources: http://www.kde-apps.org/content/show.php/QtiPlot?content=14826

2009/06/22 - QtiPlot 0.9.7.8 released. Special thanks to Olivier Dubreuil for his bug reports and suggestions. Changes:

* Slightly modified the Qwt 5.2 library in order to allow painting of 2D curves using floating point coordinates, resulting in a more accurate/smoother represantation of the data (the modified source code of this library is shipped with the QtiPlot archive in the "3rdparty/qwt" folder).
* Implemented waterfall plots.
* Added a powerfull exploration tool for 2D plots combining panning and continuous zooming.
* Improved the export of 2D plots to the EMF format on Windows. Now when you copy a plot window or a layer, the images are stored to the clipboard as metafiles using EmfEngine.
* Improved contour line plots. Now the user has the possibility to:
o use the formula defined for the data source matrix in order to calculate Z values, resulting in smoother contour lines
o display text labels for the Z contour levels
o define custom Z levels
o define custom colors and pen widths for each Z level
* Improved Python scripting for 2D curves. Added scripting support for:
o image/contour line plots (spectrograms)
o histograms
o waterfall plots
* Fixed crash when calling newNote() in a Python script with Python 2.6 (thanks to Knut Franke).
* Improved color handling for curves/symbols in a 2D plot:
o the user can define custom colors.
o it is straightforward to restrict all colors to a gray scale or to define indexed colors via the "Palette" plot layer context menu (Feature Request #00381).
* Improved data fitting capabilities:
o fixed fitting with user defined functions containing removable singularities.
o improved display range of prediction and confidence bands.
o implemented Feature Request #004587: "Fit through zero".
o added planck_wavelength fit plugin, thanks to Knut Franke.
* Improved the display of analytic 2D functions: in case of logarithmic x-scales the values are calculated using equidistant data points on the log scale.
* Improved 2D plot layers context menu.
* Fixed file saving for filenames containing dot signes.
* Added Chinese translation thanks to Fox (Chen Zhang).

[Artur Rona (ari-tczew)]

 qtiplot (0.9.7.8-1) unstable; urgency=low

   * New upstream release. (Closes: #504646, #527650, #506929,
     #539001, #532480)

   [ Fathi Boudra ]
   * Remove myself from Uploaders field.

   [ Gudjon I. Gudjonsson ]
   * Fix the watch file
   * Add get-orig-source target
   * Remove patch 02_docbook_stylesheet.diff, not needed
   * Change dependency from liborigin-dev to liborigin2-dev
   * Remove dependency on qwtplot3d since it is statically linked in
     in this version
   * Add patch 02_sip_fix.diff with changes from upstream,
     fixes compiling.
   * Remove dh_desktop
   * Bump standards version to 3.8.3, no changes needed
   * Add README.source
   * Add dependency on libpng12-dev
   * Remove patch 04_add_includes.diff. Fixed in upstream
   * Remove muparser and liborigin from copyright file
   * Add qwtplot3d to copyright file
   * Fix GPL versions in copyright file
   * Fix patch 03_fix_qti_wordlist.txt_path.diff (Closes: #475522)
   * Add rm manual/html/qtiplot.adp to clean target
   * Add doc-base file

 -- Gudjon I. Gudjonsson <email address hidden> Tue, 08 Sep 2009 08:47:51 +0200

[Scott Howard (showard314)]

Some problems need to be fixed before the sync [1] (see below). We can wait for debian do fix it, or we could do a quick Ubuntu patch and then sync when Debian gets it. If I have time later, I'll try to get a ppa up and a merge proposal for Ubuntu with the patch.

[1] http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=547273

Package: qtiplot
Version: 0.9.7.8-1
User: <email address hidden>
Usertags: ftbfs-gcc-4.4
Tags: patch

Your package fails to build with GCC 4.4 for two reasons:

1. GCC 4.4 has cleaned up some more C++ headers. You always have to
#include headers directly and cannot rely for things to be included
indirectly.

2. you're using a #elif without any condition when you really want a
#else.

You can reproduce this problem with gcc-4.4/g++-4.4 from unstable.

> Automatic build of qtiplot_0.9.7.8-1 on em64t by sbuild/amd64 0.53
...
> g++ -c -pipe -fno-exceptions -g -fPIC -D_REENTRANT -Wall -W -DGL2PS_HAVE_LIBPNG -DGL2PS_HAVE_ZLIB -DQT_OPENGL_LIB -DQT_GUI_LIB -DQT_CORE_LIB -DQT_SHARED -I/usr/share/qt4/mkspecs/linux-g++ -I. -I/usr/include/qt4/QtCore -I/usr/include/qt4/QtGui -I/usr/include/qt4/QtOpenGL -I/usr/include/qt4 -Iinclude -I/usr/X11R6/include -Itmp -o tmp/qwt3d_function.o src/qwt3d_function.cpp
> src/qwt3d_function.cpp: In member function 'virtual bool Qwt3D::Function::create()':
> src/qwt3d_function.cpp:80: error: 'stderr' was not declared in this scope
> src/qwt3d_function.cpp:80: error: 'fprintf' was not declared in this scope
> make[2]: *** [tmp/qwt3d_function.o] Error 1
> make[2]: Leaving directory `/build/tbm/qtiplot-0.9.7.8/3rdparty/qwtplot3d'
> make[1]: *** [sub-3rdparty-qwtplot3d-make_default] Error 2
> make[1]: Leaving directory `/build/tbm/qtiplot-0.9.7.8'

and:

> src/scripting/ScriptEdit.cpp:295:6: error: #elif with no expression

--- qtiplot/src/scripting/ScriptEdit.cpp~ 2009-09-18 08:15:25.000000000 +0000
+++ qtiplot/src/scripting/ScriptEdit.cpp 2009-09-18 08:15:34.000000000 +0000
@@ -292,7 +292,7 @@
  console->setPlainText(err);
  if (!console->isVisible())
   ((QDockWidget *)console->parent())->show();
-#elif
+#else
  int start = printCursor.position();
  printCursor.insertText(err);
  printCursor.setPosition(start, QTextCursor::KeepAnchor);
--- 3rdparty/qwtplot3d/src/qwt3d_function.cpp~ 2009-09-18 07:50:47.000000000 +0000
+++ 3rdparty/qwtplot3d/src/qwt3d_function.cpp 2009-09-18 07:51:11.000000000 +0000
@@ -1,5 +1,6 @@
 #include "qwt3d_surfaceplot.h"
 #include "qwt3d_function.h"
+#include <cstdio>

 using namespace Qwt3D;

--
Martin Michlmayr
http://www.cyrius.com/

[Scott Howard (showard314)]

the above patch (linked above) fixes the FTBFS, however a new one is introduced:

src/plot2D/ArrowMarker.cpp: In member function 'virtual void ArrowMarker::draw(QPainter*, const QwtScaleMap&, const QwtScaleMap&, const QRect&) const':
src/plot2D/ArrowMarker.cpp:64: error: 'scaledPen' is not a member of 'QwtPainter'

which is related to:
https://bugs.launchpad.net/ubuntu/+source/qwt/+bug/395608

scaledPen is included in qwt 5.2, but not 5.1 apparently. We'll need to get a freeze exemption for qwt (sync with debian) then try to build to the linked bzr branch.

[Scott Howard (showard314)]

qwt sync request is at:
https://bugs.launchpad.net/ubuntu/+source/qwt/+bug/433718

[Scott Howard (showard314)]

qwt was synced with debian, this package now builds.

See attached build log, and it can be installed from the following PPA for anyone wanting to take it for a spin before it hits the repos:

https://launchpad.net/~showard314/+archive/ppa

@MOTU sponsors: no API changes, just bug fixes and coding improvements. See linked merge proposal for sync with debian and ubuntu change (FTBFS patch)

[James Westby (james-w)]

Taking for sponsorship from the merge request.

Thanks,

James

[Scott Howard (showard314)]

Since this version has some new features, we are going to need a feature freeze exception. To get the feature freeze exception, we need thorough testing of the new package.

Could some of the subscribers of this bug install qtiplot from:
https://launchpad.net/~showard314/+archive/ppa

Please test that package with some of your current scripts and data, and reply here if it is working well (or with any problems). Once it is tested we can apply for the FFE and get this into Karmic.

Also, I'd like to thank James Westby for helping with the sponsorship.

[Scott Howard (showard314)]

For FFE: see attached install log and screenshot of the package installed from ppa. I did some basic testing of the program and did not have any problems. The build log and changelog (debian and upstream) are located above as well.

More testing from other users would be appreciated.

[Dmitry Tantsur (divius)]

2009/09/14 - QtiPlot 0.9.7.9 released. Maybe test it an a ppa?

[Scott Howard (showard314)]

Thanks for the comment. The new version 0.9.7.9 requires a custom qwt
library. In Ubuntu and Debian, the standard library is provided by the
package libqwt5-qt4-dev. For the most part, we don't want to mess with
custom libraries this close to release (beta freeze started this
morning and we've been in feature freeze for a few weeks). However,
the changes in the custom library have been accepted by the upstream
authors of qwt in their svn.

In order to get qtiplot 0.9.7.9 into Karmic, we would need to get a
freeze exception to get a new qwt into karmic (after testing it) and
then get a freeze exception for qtiplot 0.9.7.9 (after testing it).
This seems difficult with the remaining time in the release schedule,
and the motu-release team might not want to try it. That's why I think
it would be best if we could get qtiplot 0.9.7.8 into Karmic and then
sync with Debian after the release in order to get the new
libqwt5-qt4-dev package and qtiplot version 0.9.7.9.

However, we still would need a freeze exception for 0.9.7.8, otherwise
we'll ship with the same version that was in Jaunty.

2009/9/24 Дмитрий "Divius" Танцур <email address hidden>:
> 2009/09/14 - QtiPlot 0.9.7.9 released. Maybe test it an a ppa?
>
> --
> New upstream release qtiplot 0.9.7.8 (sync with debian and fix for debian's FTBFS)
> https://bugs.launchpad.net/bugs/374543
> You received this bug notification because you are a direct subscriber
> of the bug.
>
> Status in “qtiplot” package in Ubuntu: New
> Status in “qtiplot” package in Debian: New
>
> Bug description:
> Sources: http://www.kde-apps.org/content/show.php/QtiPlot?content=14826
>
> 2009/06/22 - QtiPlot 0.9.7.8 released. Special thanks to Olivier Dubreuil for his bug reports and suggestions. Changes:
>
> * Slightly modified the Qwt 5.2 library in order to allow painting of 2D curves using floating point coordinates, resulting in a more accurate/smoother represantation of the data (the modified source code of this library is shipped with the QtiPlot archive in the "3rdparty/qwt" folder).
> * Implemented waterfall plots.
> * Added a powerfull exploration tool for 2D plots combining panning and continuous zooming.
> * Improved the export of 2D plots to the EMF format on Windows. Now when you copy a plot window or a layer, the images are stored to the clipboard as metafiles using EmfEngine.
> * Improved contour line plots. Now the user has the possibility to:
> o use the formula defined for the data source matrix in order to calculate Z values, resulting in smoother contour lines
> o display text labels for the Z contour levels
> o define custom Z levels
> o define custom colors and pen widths for each Z level
> * Improved Python scripting for 2D curves. Added scripting support for:
> o image/contour line plots (spectrograms)
> o histograms
> o waterfall plots
> * Fixed crash when calling newNote() in a Python script with Python 2.6 (thanks to Knut Franke).
> * Improved color handling for curves/symbols in a 2D plot:
> o the user can define custom colors.
> o it is straightforward to restrict all colors to a gray scale or to define indexed colors via the "Palette" plot layer context menu (Feature Request #00381).
> * Impr...

[StefanPotyra (sistpoty)]

yes, that makes sense. +1 for 0.9.7.8.

[ASteppke (asteppke-gmail)]

Hi,

I tried the new 0.9.7.8 version from https://launchpad.net/~showard314/+archive/ppa under Ubuntu 9.10.

While the installation and basic operations are fine Python scripting support is either broken or missing.

To reproduce the bug try the following steps:
1. start qtiplot
2. go in the menu to Scripting->Scripting-Language
3. Select Python and press ok.
4. Error message: "Scripting language Python failed to initialize", additionally I get under Kubuntu the message "Sorry, QtiPlot closed unexpectedly." but the program still seems to run after I close the error message.

Thanks for looking into this!
Alexander

[ASteppke (asteppke-gmail)]

Sorry, just found out that the Python bug is already fixed in the upstream SVN release. So manually fixing it with the information from Bug #435918 resolves the problem with the 0.9.7.8 version from https://launchpad.net/~showard314/+archive/ppa under Ubuntu 9.10.

[Scott Howard (showard314)]

Thanks Steppke, I tried to include upstream's fix for that specific
into my ppa - but it would not work (see ~ppa2 and ~ppa3). I included
the patch, the patch is applied, but scripting still does not work. If
anyone else wants to take a stab at it, I'd appreciate it. That
specific bug has been around in Ubuntu for atleast two releases.

On Thu, Oct 1, 2009 at 2:25 PM, ASteppke <email address hidden> wrote:
> Sorry, just found out that the Python bug is already fixed in the
> upstream SVN release. So manually fixing it with the information from
> Bug #435918 resolves the problem with the 0.9.7.8 version from
> https://launchpad.net/~showard314/+archive/ppa under Ubuntu 9.10.
>
> --
> New upstream release qtiplot 0.9.7.8 (sync with debian and fix for debian's FTBFS)
> https://bugs.launchpad.net/bugs/374543
> You received this bug notification because you are a direct subscriber
> of the bug.
>
> Status in “qtiplot” package in Ubuntu: New
> Status in “qtiplot” package in Debian: New
>
> Bug description:
> Sources: http://www.kde-apps.org/content/show.php/QtiPlot?content=14826
>
> 2009/06/22 - QtiPlot 0.9.7.8 released. Special thanks to Olivier Dubreuil for his bug reports and suggestions. Changes:
>
> * Slightly modified the Qwt 5.2 library in order to allow painting of 2D curves using floating point coordinates, resulting in a more accurate/smoother represantation of the data (the modified source code of this library is shipped with the QtiPlot archive in the "3rdparty/qwt" folder).
> * Implemented waterfall plots.
> * Added a powerfull exploration tool for 2D plots combining panning and continuous zooming.
> * Improved the export of 2D plots to the EMF format on Windows. Now when you copy a plot window or a layer, the images are stored to the clipboard as metafiles using EmfEngine.
> * Improved contour line plots. Now the user has the possibility to:
> o use the formula defined for the data source matrix in order to calculate Z values, resulting in smoother contour lines
> o display text labels for the Z contour levels
> o define custom Z levels
> o define custom colors and pen widths for each Z level
> * Improved Python scripting for 2D curves. Added scripting support for:
> o image/contour line plots (spectrograms)
> o histograms
> o waterfall plots
> * Fixed crash when calling newNote() in a Python script with Python 2.6 (thanks to Knut Franke).
> * Improved color handling for curves/symbols in a 2D plot:
> o the user can define custom colors.
> o it is straightforward to restrict all colors to a gray scale or to define indexed colors via the "Palette" plot layer context menu (Feature Request #00381).
> * Improved data fitting capabilities:
> o fixed fitting with user defined functions containing removable singularities.
> o improved display range of prediction and confidence bands.
> o implemented Feature Request #004587: "Fit through zero".
> o added planck_wavelength fit plugin, thanks to Knut Franke.
> * Improved the display of analytic 2D functions: in case of logarithmic x-scales the values are calculated using equidistant data points on the log scale.
> * Improved 2D plot layers context menu.
> * Fixed file saving for filenames contain...

[Scott Kitterman (kitterman)]

I see a huge red flag on this one (from Debian changelog):

   * Remove dependency on qwtplot3d since it is statically linked in
     in this version

NACK from me for introducing code copies into the archive.

[Scott Kitterman (kitterman)]

http://bugs.gentoo.org/show_bug.cgi?id=265623 has some hints on the qwt api change. I think either teach the version we have now to talk to the new qwt or teach 0.9.7.8 to use an external qwtplot3d again (bonus points for doing the same with liborigin and zlib).

[Scott Howard (showard314)]

Thanks for the pointers and help! The author of qtiplot does contribute alot of upstream changes and improvements to those libraries, but unfortunately for us, qtiplot is released prior to the changed libraries being released, causing their need for including statically linked libraries. It seems that he has added new functionality to qwtplot3d that we don't have in our version yet, and we won't get until the archives open for syncing. On top of that, they have released up to 0.9.7.10, and those versions include all the statically linked libraries in 0.9.7.8 and a statically linked qwt!

For now, I'll have people use the PPA as a work around for bugs [1], and work on qtiplot in Lucid.

[1] https://bugs.launchpad.net/debian/+bug/435476

[Fabrice Coutadeur (fabricesp)]

Hi,

As it seems it's a no go for the sync in Karmic, I'll unsubscribe u-u-s.
Fell free to subscribe Ubuntu Sponsors for Universe again is needed.

Thanks,
Fabrice

[Leandro (leandromartinez98)]

I considering the use of the new qtiplot version in Karmic, this bug must be
taken into account:

https://bugs.launchpad.net/ubuntu/+source/qtiplot/+bug/435476

This bug makes the current version delivered totally useless. If the new version
will not arrive to Karmic, then some previous stable version should be
released, otherwise qtiplot will be totally broken.

[StefanPotyra (sistpoty)]

even though we might fix up qtiplot to use the shared library via an sru, the new debian version just doesn't build (to be exact, it doesn't build via the included library of qwt3d). I'm not too sure if I'll get anywhere close to fixing this before FinalFreeze for universe :(.

[Scott Howard (showard314)]

Stefan, if you are working on this, see the linked branch and the PPA I linked above. I was able to get it to build using the included qwt3d. Maybe you could use that as a starting point if it helps.

[Scott Howard (showard314)]

Actually, all you need is this patch to fix the FTBFS, it's from the upstream svn (links are in the comments in the header).

[Scott Howard (showard314)]

Excuse me, this actually is from the debian bug report I linked above, it's not in SVN yet.

[StefanPotyra (sistpoty)]

Scott, thanks for the fix, however it still fails to build (as seen in debian bug 551222). I'll see what I can do to fix it, but it's extremely late in my timezone and I'm not 100% sure if I can it fixed yet.

[Scott Kitterman (kitterman)]

I think it is too late for this now as we are past final freeze for Universe and only looking for high priority bug fixes. This will get automatically sync'ed for Lucid, so marking the bug wontfix for karmic.