| Summary: | sci-visualization/scigraphica-2.1 version bump | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Jan Kohnert <bughunter> |
| Component: | New packages | Assignee: | Gentoo Science Related Packages <sci> |
| Status: | RESOLVED FIXED | ||
| Severity: | enhancement | CC: | basic, cloos, frp.bissey, grozin, jlp.bugs, pdenapo, smustudent1, ysharma |
| Priority: | High | ||
| Version: | unspecified | ||
| Hardware: | All | ||
| OS: | All | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
| Bug Depends on: | 102872 | ||
| Bug Blocks: | 154102 | ||
| Attachments: |
libscigraphica-2.0.0.ebuild
scigraphica-2.0.0.ebuild libscigraphica-2.0.0.ebuild (fixed description) diff between my and Pablos libscigraphica ebuild diff between my and Pablos scigraphica ebuild Patch to get the correct version of sg in the "scigraphuica --version" dialog Ebuild for new libscigraphica-2.1.1 Ebuild for new scigraphica-2.1.0 New configure.in.patch for sg-version 2.1.0 Ebuild for pygtk as needed by sg-2.1.0 libscigraphica-configure.patch scigraphica-arrayobject.h.patch |
||
|
Description
Jan Kohnert
2005-05-18 12:48:03 UTC
SORRY: I meant media-gfx/scigraphica. Changed summary. Best regards Jan This package depends on x11-libs/gtk+extra-2.0.0 I'm contributing an ebuild for this library (see bug #65354 ) I'll see if I can create an ebuild for scigraphica also I think that he curren classification in portage is not correct, as it is clearly a scientific application (shouldn't be in sci-mathematics or in sci-misc?) Created attachment 59426 [details]
libscigraphica-2.0.0.ebuild
An ebuild for libscigraphica
Created attachment 59427 [details]
scigraphica-2.0.0.ebuild
an ebuild for scigraphica-2.0.0
I'm submitting an ebuild for libscigraphica and scigraphica. They DON'T WORK right now, they are just an starting point (but I don't want you to make the work twice) Actually scigraphica and libscigraphica are built by these ebuilds, however scigraphica is not able to find the python modules It display the following messages when I run it in an xterm: Importing python module: pickle Importing python module: gtk ImportError: No module named gtk Importing python module: Numeric ImportError: No module named Numeric Importing python module: sys Importing python module: pysga ImportError: No module named pysga Importing python module: os any help is wellcome! I have tried to figure out the dependencies by reading the configure.in files but I'm not so sure about them. Please check. One final note: the gnome and bonobo support seem to be gone in this release. I also created all tree ebuilds and after deleting the old .scigraphica dir it seems to work. But I also get an error: It claims it cannot find numarray... I can try to make some diffs between yours and mine files if you like. Best regards Jan Just changed dependencies of libscigraphica from dev-python/numeric to dev/python/numarray. I installed numarray and it works like a charm. If I should upload my files or create diffs just let me know. @Pablo: Try to rm/rename your old scigraphica config-dir, this should do the trick. Best regards Jan After deleting the .scigraphica dir the ebuilds did work. Thanks! I'm not having problems with loading numarray (the ebuilds actually depend of numarray) Jan: It would be nice if you could check if there are any differencies between your ebuilds and mine, so that we fix any remaning problem. I think that we should try to merge our work. One thing that obviusly should be changed is the descripci After deleting the .scigraphica dir the ebuilds did work. Thanks! I'm not having problems with loading numarray (the ebuilds actually depend of numarray) Jan: It would be nice if you could check if there are any differencies between your ebuilds and mine, so that we fix any remaning problem. I think that we should try to merge our work. One thing that obviusly should be changed is the descripción of libscigraphica in my ebuild. (I forgot that) A problem that remains is that scigraphica segfaults when exiting! (this seems to be a bug in the program) Created attachment 59457 [details]
libscigraphica-2.0.0.ebuild (fixed description)
Just changed the description
> A problem that remains is that scigraphica segfaults when exiting!
> (this seems to be a bug in the program)
It does not for me. Could you please provide any kind of backtrace? I could
report it upstream to sg-devel, as I read the list since many years.
There seems also to be a bug with pango-UTF8 things and I don't seem to be able
to open old sg-files (maybe that two things are related, I will investigate).
Another problem is, that the configure-script does not seem to be able to
detect numarray correctly. No problem at our site, but I will look at this and
report upstream.
Best regards Jan
> Jan: It would be nice if you could check if there are any differencies > between your ebuilds and mine, so that we fix any remaning > problem. I think that we should try to merge our work. Just created some diffs. SRC_URI differs, but if your SRC_URI works, it' surely better than mine, as I just used a mirror at my local region. > One thing that obviusly should be changed is the descripci > Jan: It would be nice if you could check if there are any differencies > between your ebuilds and mine, so that we fix any remaning > problem. I think that we should try to merge our work. Just created some diffs. SRC_URI differs, but if your SRC_URI works, it' surely better than mine, as I just used a mirror at my local region. > One thing that obviusly should be changed is the descripción > of libscigraphica in my ebuild. > (I forgot that) That's great, also forgot that. AFAIK there is no active gnome support at this time, so I think, the IUSE="gnome" could be changed my way. One thing I changed was DEPENDS of scigraphica, as all dependencies are cached by the lib, the only thing remains in sg, is the lib itself (but better with =, as in your ebuild). Also, is there really a dependency on imlib and intltool? It seems configure would not check it. I also have a scr_unpack() funktion, but I think this is an error at my site (those were the first ebuilds I wrote). The last thing is I have pkg_postinstall with the warning to delete the old configuration, as this would be surely the first bugreport on that package. I'll send the diffs immediatly. Best reagdars Jan Created attachment 59465 [details, diff]
diff between my and Pablos libscigraphica ebuild
Created attachment 59466 [details, diff]
diff between my and Pablos scigraphica ebuild
Anybody else seen a complaint from arrayobject.h not found? It's a library in numeric - I had to symlink it to /usr/include/arrayobject.h before the build process would find it. In /usr/include/python2.3/Numeric/ it apparently isn't spotted? Nuts - libscigraphica build bottomed out again. Anybody have any ideas?
In file included from /usr/include/python2.3/Python.h:8,
from sg_python_expr.c:26:
/usr/include/python2.3/pyconfig.h:856:1: warning: "_POSIX_C_SOURCE" redefined
In file included from /usr/include/limits.h:26,
from
/usr/lib/gcc/i686-pc-linux-gnu/3.4.3-20050110/include/limits.h:122,
from
/usr/lib/gcc/i686-pc-linux-gnu/3.4.3-20050110/include/syslimits.h:7,
from
/usr/lib/gcc/i686-pc-linux-gnu/3.4.3-20050110/include/limits.h:11,
from /usr/lib/glib-2.0/include/glibconfig.h:11,
from /usr/include/glib-2.0/glib/gtypes.h:30,
from /usr/include/glib-2.0/glib/galloca.h:30,
from /usr/include/glib-2.0/glib.h:30,
from /usr/include/gtk-2.0/gdk/gdktypes.h:32,
from /usr/include/gtk-2.0/gdk/gdkcolor.h:4,
from /usr/include/gtk-2.0/gdk/gdk.h:30,
from sg_python_expr.c:19:
/usr/include/features.h:190:1: warning: this is the location of the previous
definition
sg_python_expr.c: In function `sg_eval_expr_double':
sg_python_expr.c:133: warning: unused variable `parray'
sg_python_expr.c:134: warning: unused variable `apt'
sg_python_expr.c:136: warning: unused variable `retarray'
sg_python_expr.c: At top level:
sg_python_expr.c:130: warning: unused parameter 'expect'
sg_python_expr.c: In function `sg_eval_expr_double':
sg_python_expr.c:213: warning: control reaches end of non-void function
sg_python_expr.c: At top level:
sg_python_expr.c:214: warning: type defaults to `int' in declaration of `len'
sg_python_expr.c:214: error: `expect' undeclared here (not in a function)
sg_python_expr.c:214: error: `array' undeclared here (not in a function)
sg_python_expr.c:214: warning: data definition has no type or storage class
sg_python_expr.c:215: error: parse error before "if"
sg_python_expr.c:218: error: parse error before '&' token
sg_python_expr.c:218: warning: type defaults to `int' in declaration of
`g_array_append_vals'
sg_python_expr.c:218: error: conflicting types for 'g_array_append_vals'
/usr/include/glib-2.0/glib/garray.h:77: error: previous declaration of
'g_array_append_vals' was here
sg_python_expr.c:218: error: conflicting types for 'g_array_append_vals'
/usr/include/glib-2.0/glib/garray.h:77: error: previous declaration of
'g_array_append_vals' was here
sg_python_expr.c:220: warning: type defaults to `int' in declaration of `retarray'
sg_python_expr.c:220: warning: data definition has no type or storage class
sg_python_expr.c:221: warning: type defaults to `int' in declaration of `num'
sg_python_expr.c:221: warning: data definition has no type or storage class
sg_python_expr.c:222: error: parse error before '(' token
make[3]: *** [sg_python_expr.lo] Error 1
make[3]: Leaving directory
`/var/tmp/portage/libscigraphica-2.0.0/work/libscigraphica-2.0.0/scigraphica'
make[2]: *** [all-recursive] Error 1
make[2]: Leaving directory
`/var/tmp/portage/libscigraphica-2.0.0/work/libscigraphica-2.0.0/scigraphica'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory
`/var/tmp/portage/libscigraphica-2.0.0/work/libscigraphica-2.0.0'
make: *** [all] Error 2
(In reply to comment #15) > Anybody else seen a complaint from arrayobject.h not found? It's a library in > numeric - I had to symlink it to /usr/include/arrayobject.h before the build > process would find it. In /usr/include/python2.3/Numeric/ it apparently isn't > spotted? Look in the archives of scigraphica-devel mailing list. Adrian had some problems, because newer versions of numarray weren't catched correctly; also your other described error could be spotted by this. Adrian changed depedency back to Numeric. Are you sure to use Numeric instead of nummarray? Maybe having Numeric *and* nummarry installed? Waiting for comments, I will report upstream. Best regards Jan Created attachment 61171 [details, diff]
Patch to get the correct version of sg in the "scigraphuica --version" dialog
This patch fixes a small error in configure.in, which precents sg to give the
correct version with the command "scigraphica --version"
It is tested on my system and worked.
Just applied it and afterwards
libtoolize --force && aclocal && autoconf && autoheader && automake
Then the well known ./configure && make && make install ;)
Best regards Jan
How are these ebuilds working? Are they about ready to go in, for 2.1.0? > How are these ebuilds working? Are they about ready to go in, for 2.1.0?
I'll have a look at it ASAP, also concerning 2.1.0, hopefully get new finished
ebuilds for minimum ~x86 at the end of this week...
Best regards Jan
Error on the 2.1.0 side when installing libscigraphica:
../../pixmaps -I../../scigraphica -I../../scigraphica/dialogs -DXTHREADS
-D_REENTRANT -DXUSE_MTSAFE_API -I/usr/include/gtk-2.0 -I/usr/lib/gtk-2.0/include
-I/usr/include/atk-1.0 -I/usr/include/pango-1.0 -I/usr/include/freetype2
-I/usr/include/freetype2/config -I/usr/include/glib-2.0
-I/usr/lib/glib-2.0/include -DXTHREADS -D_REENTRANT -DXUSE_MTSAFE_API
-I/usr/include/gtkextra-2.0 -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include
-I/usr/include/gtk-2.0 -I/usr/lib/gtk-2.0/include -I/usr/include/atk-1.0
-I/usr/include/pango-1.0 -I/usr/include/freetype2
-I/usr/include/freetype2/config -I/usr/include/python2.4 -march=athlon-xp
-O3 -pipe -g -O -W -Wall -DWITH_WARNINGS -MT sg_entry.lo -MD -MP -MF
".deps/sg_entry.Tpo" \
-c -o sg_entry.lo `test -f 'sg_entry.c' || echo './'`sg_entry.c; \
then mv -f ".deps/sg_entry.Tpo" ".deps/sg_entry.Plo"; \
else rm -f ".deps/sg_entry.Tpo"; exit 1; \
fi
In file included from ../../scigraphica/sg.h:44,
from sg_dialog.c:21:
../../scigraphica/sg_python.h:35:25: arrayobject.h: No such file or directory
sg_dialog.c:48: warning: unused parameter 'dialog'
sg_dialog.c:56: warning: unused parameter 'dialog'
sg_dialog.c:57: warning: unused parameter 'event'
sg_dialog.c:68: warning: unused parameter 'dialog'
sg_dialog.c:248: warning: unused parameter 'button'
sg_dialog.c:256: warning: unused parameter 'button'
sg_dialog.c:264: warning: unused parameter 'button'
sg_dialog.c:272: warning: unused parameter 'button'
sg_dialog.c:280: warning: unused parameter 'button'
sg_dialog.c:273: warning: 'button_clicked_apply' defined but not used
make[3]: *** [sg_dialog.lo] Error 1
make[3]: *** Waiting for unfinished jobs....
I have numeric installed, so I don't know quite what's going on here...
Oh, I see - for some weird reason the include file are in python2.3??? This is a brand new gentoo install, how could that have happened? Hi, I have tried the scigraphica ebuild that is in portage and I am quite interested in testing an ebuild for the new version (0.8.0 segfault on me each time I tried to plot a 2nd 3D graph). I can test for both ppc (my work machine) and x86 (home machine). Where could I get an ebuild without dealing with working which patch goes where and which one should be applied from the bug list? Hi Francois,
Since it's not in Portage yet, you need to download the ebuilds from Bugzilla
and place them in a Portage overlay on your system, applying all the necessary
patches. See the Gentoo Handbook for more details.
(In reply to comment #24) > Hi Francois, > > Since it's not in Portage yet, you need to download the ebuilds from Bugzilla > and place them in a Portage overlay on your system, applying all the necessary > patches. See the Gentoo Handbook for more details. Yes if it had been in portage I would have had a go at it already, I was hoping to bypass the patching pain :p (to me), and working out which patches are the necessary ones. I guess I will give it a go on Monday. (In reply to comment #2) > This package depends on x11-libs/gtk+extra-2.0.0 > I'm contributing an ebuild for this library > (see bug #65354 ) > I'll see if I can create an ebuild for scigraphica > also This is a bug for an ebuild of pygsl not gtk+extra. The correct bug is #93394 :) . Should we make the present bug depend on your gtk+extra bug Pablo? Sorry for the last response: Much too less time in the last weeks. ;) I created new ebuild for the new versions: gtkextra-2.1.1, libscigraphica-2.1.1 and scigraphica-2.1.0 However, there is still one known problem, which I worked around building a new ebuild pygtk-2.6.1-r1 using numarray instead of numeric. The ebuilds all work fine (at least on my system), but I don't know if there are any side effects by the new pygtk ebuild or if numarray is even used by the new ebuild. BTW: Could somebody point this one to the pygtk maintainer? It is needed to be checked. Best regards Jan Created attachment 66151 [details]
Ebuild for new libscigraphica-2.1.1
Created attachment 66152 [details]
Ebuild for new scigraphica-2.1.0
Created attachment 66153 [details, diff]
New configure.in.patch for sg-version 2.1.0
Created attachment 66154 [details]
Ebuild for pygtk as needed by sg-2.1.0
This ebuild is needed by scigraphica-2.1.0.ebuild, could somebody please point
this to the correct maintainer to check it?
I short:
pygtk needs numeric but scigraphica depends on numarray and produces strange
build errors if numeric or both packages are installed; but at the same time sg
depends on pygtk...
Just created a new bug-report with the question concerning pygtk. See: http://bugs.gentoo.org/show_bug.cgi?id=102872 Hopefully somebody can clearify this one... Best regards Jan Ran fine but this came at exit: (scigraphica:8914): GLib-GObject-WARNING **: invalid uninstantiatable type `glong' in cast to `GtkWidget' (scigraphica:8914): Gtk-CRITICAL **: gtk_container_remove: assertion `GTK_IS_WIDGET (widget)' failed Segmentation fault Added scigraphica-2.1.0 and libscigraphica-2.1.1 in the gentoo scientific overlay at http://gentooscience.org. Only slight modifications from the ebuilds proposed here. pygtk was not modified to accomodate numarray, since I have not encountered any problem yet with coexistence of numarray for scigraphica and Numeric for pygtk. I do not used scigraphica, so please test and report in this bug. Created attachment 87206 [details, diff]
libscigraphica-configure.patch
Here's the patch to libscigraphica to fix the case when numarray and numeric are simultaneously installed.
Created attachment 87207 [details, diff]
scigraphica-arrayobject.h.patch
Here is the equivalent patch to scigraphica.
The ebulds scigraphica-2.1.0 and libscigraphica-2.1.1 (from gentooscience.org) install OK. However, scigraphics always segfaults when I exit from it. I updated the scigraphica ebuilds in the overlay (see comment #34). It now compiles with gcc-41, and I applied other patches, among which are inspired from spyderous overlay. It still segfaults when exiting (and in other cases as seen in the scigraphica mailing lists). Fixing it does not look obvious. Finally in the main tree, thank you all for your patience! |
