## scigraphica-devel — Development discussion

You can subscribe to this list here.

 2000 2001 2002 2003 2004 2005 2006 2008 Jan Feb Mar Apr May (27) Jun (38) Jul (16) Aug (69) Sep (31) Oct (26) Nov (72) Dec (13) Jan (9) Feb (222) Mar (216) Apr (114) May (19) Jun (70) Jul (246) Aug (182) Sep (99) Oct (300) Nov (213) Dec (190) Jan (51) Feb (19) Mar (14) Apr (5) May (41) Jun (12) Jul (9) Aug (16) Sep (40) Oct (10) Nov (28) Dec (8) Jan (13) Feb (3) Mar (31) Apr (12) May (5) Jun (13) Jul (4) Aug (1) Sep Oct (2) Nov (3) Dec (2) Jan (2) Feb (1) Mar Apr (19) May (4) Jun Jul Aug (1) Sep (8) Oct (1) Nov (13) Dec (3) Jan (29) Feb (10) Mar Apr (3) May (36) Jun (19) Jul (5) Aug (11) Sep (3) Oct (12) Nov (16) Dec Jan (1) Feb (4) Mar (2) Apr May (1) Jun (10) Jul (6) Aug (1) Sep (2) Oct Nov Dec Jan Feb Mar Apr May Jun (1) Jul Aug Sep Oct Nov Dec
S M T W T F S

1

2
(3)
3
(10)
4
(14)
5
(16)
6
(2)
7
(14)
8
(3)
9

10
(9)
11
(10)
12
(10)
13
(9)
14
(6)
15
(7)
16
(3)
17
(19)
18
(16)
19
(14)
20
(9)
21
(5)
22
(4)
23
(1)
24

25
(1)
26
(2)
27

28
(1)
29

30

31
(2)

Showing results of 190

1 2 3 .. 8 > >> (Page 1 of 8)
 [Scigraphica-devel] Any doubt about Scigraphica, Erros in operations. From: Ronaldo Reis Jr. - 2001-12-31 21:41:45 ```Hi all, I have followed the scigraphica since 0.5x version. Now, in version 0.8 it is very stable than others. It is my future "Linux Sigma Plot" :)))). But I have any questions about their use. Fisrt I create a sheet with simple data. X Y 1 9 2 21 3 29 4 42 5 48 6 61 7 69 8 82 9 90 10 99 The graph is great, but I need to plot an function in this points to create an ajusted curve. The function is f1(x)= -0.06667 + 10.01212*X. This function make a linear fitted curve for the points. I make it with the Function, but not work:(((( Then, I try to make the ajusted point in the sheet, dont work too. Is very strange the behavior, I try make an column that values is: 1.5+2.5 and the returned value is 3.0. I try some others operations and all return error in operation with decimal. It is a Bug??? Thanks for all and congratulations for the program. Ronaldo -- Never trust an operating system. -- | //|\\ [*****************************][*******************] || ( õ õ ) [Ronaldo Reis Júnior ][PentiumIII-600 ] | V [ESALQ/USP-Entomologia, CP-09 ][HD: 30 + 10 Gb ] || / l \ [13418-900 Piracicaba - SP ][RAM: 128 Mb ] | /(lin)\ [Fone: 19-429-4199 r.229 ][Video: SiS620-8Mb ] ||/(linux)\ [chrysopa@... ][Modem: Pctel-onboar] |/ (linux) \[ICQ#: 5692561 ][SO: CL 7.0 (2.2.19)] || ( x ) [*****************************][*******************] ||| _/ \_Powered by Conectiva Linux 7.0 D+:) | Lxuser#: 205366 ```
 [Scigraphica-devel] Statusbar implemented From: R. Lahaye - 2001-12-31 08:59:55 ```Adrian Feiguin wrote: > > On Tue, 18 Dec 2001, R. Lahaye wrote: > > #else > > /* Do something equivalent for gtk!*/ > > #endif > > Under gnome you get messages in the bottom of the iconlist telling what > the program is doing, we have to do something equivalent for gtk. I've centralized the statusbar/progressbar stuff in two new files: src/sg_project_statusbar.[ch]. It's rather primitive yet, but it works. In CVS as of now. The generic function "sg_project_statusbar_flash(msg)" pushes the message onto the gtk/gnome statusbar. Regards, Rob ```
 [Scigraphica-devel] scigraphica bug in tab delimited ascii import From: tomi nurminen - 2001-12-28 16:54:00 ```"rename columns if 1. line contains labels" does not work properly, if there is spaces between words.. For example if first column should be "Elapsed time[s]", the first column will be named "Elapsed" and the second column "time[s]". Scigraphica also segfaults during this kind of import ```
 Re: [Scigraphica-devel] Solution for Bug #494545 From: Robert Schwebel - 2001-12-26 12:28:25 ```On Wed, 26 Dec 2001, R. Lahaye wrote: > You seemed to have problems with the db-library, but that library is > only required if you simultaneously use GNOME support for SciGraphica > (you're not using GNOME, or are you?). Nope. > So the testing for the files in the db-devel package (fatal if not > found!) should only be preformed when also GNOME support is used. > > Conrad, could you add the necessary lines for this in the configure.in > script? Please see below for other link problems with the db-library. Ok, please tell me if you have a version available for testing (probably in CVS? Or should I try a snapshot?). Robert -- +--------------------------------------------------------+ | Dipl.-Ing. Robert Schwebel | http://www.pengutronix.de | | Pengutronix - Linux Solutions for Science and Industry | | Braunschweiger Str. 79, 31134 Hildesheim, Germany | | Phone: +49-5121-28619-0 | Fax: +49-5121-28619-4 | +--------------------------------------------------------+ ```
 Re: [Scigraphica-devel] Solution for Bug #494545 From: R. Lahaye - 2001-12-26 05:31:02 ```Robert, You seemed to have problems with the db-library, but that library is only required if you simultaneously use GNOME support for SciGraphica (you're not using GNOME, or are you?). So the testing for the files in the db-devel package (fatal if not found!) should only be preformed when also GNOME support is used. Conrad, could you add the necessary lines for this in the configure.in script? Please see below for other link problems with the db-library. Regards, Rob. Robert Schwebel wrote: > > Hi John, > > On Sat, 22 Dec 2001, John Bland wrote: > > If you follow the instructions at the bottom of FAQ.compile and > > install the dependencies listed there scigraphica should compile > > without any problems on SuSE 7.2, but if you still experience problems > > there are rpms available. > > Hmm, only if there is no other way ;) I think I have now installed > everything from the list (db-devel was missing), but it still doesn't > link out of the box: > > ----------8<---------- > /usr/lib/python2.0//config/libpython2.0.a(posixmodule.o): In function > `posix_tmpnam': > posixmodule.o(.text+0x30f6): the use of `tmpnam_r' is dangerous, better > use `mkstemp' > /usr/lib/python2.0//config/libpython2.0.a(posixmodule.o): In function > `posix_tempnam': > posixmodule.o(.text+0x3012): the use of `tempnam' is dangerous, better use > `mkstemp' > /usr/lib/python2.0//config/libpython2.0.a(bsddbmodule.o): In function > `newdbhashobject': > bsddbmodule.o(.text+0x76): undefined reference to `__db185_open' > /usr/lib/python2.0//config/libpython2.0.a(bsddbmodule.o): In function > `newdbbtobject': > bsddbmodule.o(.text+0x1a3): undefined reference to `__db185_open' > /usr/lib/python2.0//config/libpython2.0.a(bsddbmodule.o): In function > `newdbrnobject': > bsddbmodule.o(.text+0x2cd): undefined reference to `__db185_open' > collect2: ld returned 1 exit status > make[3]: *** [scigraphica] Error 1 > make[3]: Leaving directory `/home/local/scigraphica-0.8.0/src' > make[2]: *** [all-recursive] Error 1 > make[2]: Leaving directory `/home/local/scigraphica-0.8.0/src' > make[1]: *** [all-recursive] Error 1 > make[1]: Leaving directory `/home/local/scigraphica-0.8.0' > make: *** [all-recursive-am] Error 2 > ---------->8---------- > > __db185_open is in /usr/lib/libdb.so: > > callisto: # nm -n /usr/lib/libdb.so |grep __db185_open > 0000f750 T __db185_open > > But the order of the linker argument seems to be wrong: -lpython2.0 comes > before -ldb. I've temporarily added '-L/usr/lib -ldb' before '-lpython2.0' > and then it compiles. The warnings are still there, but at least it works. ```
 [Scigraphica-devel] Re: build notes so far on Irix6.5 From: R. Lahaye - 2001-12-25 07:08:36 ```Mathew, I have applied several changes to the autogen/configure build in CVS. The problem is that I do not encounter your problems and thus cannot really test whether my patches actually deal with it properly. I appreciate if you could do a test and see which problems still remain. Thanks. ****************************************************************** * 1) autgen.sh fails unless macros/gtkextra is removed. I have modified autogen.sh to search in `aclocal --print-ac-dir` for gtkextra.m4 and, if found, remove macros/gtkextra.m4. That should do here. * 2) if either --with-termcap-path or --with_readline_path are not set, * configure tries to compile its test program with the flag -L. On Irix, * this causes the library search path to get deleted. I have added tests in configure.in to only add this if these paths are specified. * 3) need readline to be installed. I don't remember this being mentioned * as a requirement. The FAQ.compile mentions readline; isn't that sufficient? * 4) arrayobject.h only searched for in python system path. Should be * configurable for cases where Numeric is installed elsewhere Added a --with-python-numeric-path to the configure options, in case numeric-python is installed in an unusual directory. * 5) During config, if imlib is not found, configure says image handling * is disabled. But when zvt is built, gdk_imlib is still linked in I removed all the imlib links in sg/zvt/Makefile.am. * 6) msgfmt must be installed Yes. Added a line to FAQ.compile. * 7) in src/python includedir is passed to CFLAGS,picks * up old gtkextra include files I have changed the order of a few includes in Makefile.am's, to give lowest priority to includedir. I hope that helps. * 8) no mpc library. Had to remove -lmpc. According to the Irix man pages there * are a few parallelizing libraries but they are called libc_mp.a, * /usr/lib64/mips3/libmp.a and /usr/lib64/mips4/libmp.a I have no idea which macro is adding the -lmpc on IRIX. It's not done on my PC (FreeBSD). Do you have a clue? * 9) you can set the readline include path to anything you want as a switch * to configure. But this value is ignored. The values of --with-readline-include-path and --with-termcap-include-path are added to \$CFLAGS. That should work, or is anything wrong with that? * 10) Unresolved text symbols in src directory. On the link line, * Libraries are linked in before they * are needed. For example, the line -L../zvt -lsgvt appears AFTER libgtk is * linked in. Also, -lz needs to be added -L../zvt -lsgvt have been moved to the front of the LIBS list; added a test for "-lz" in configure.in. *********************************************************** Ideally you should now be able to compile CVS on IRIX; or at least encountering much less problems :). Please let me know if that's indeed the case. Regards, Rob. ```
 Re: [Scigraphica-devel] Solution for Bug #494545 From: R. Lahaye - 2001-12-23 02:51:09 ```Robert Schwebel wrote: > > > > Hmm, gdk_imlib is supplied by the imlib-devel package, which is listed in > > the requirements in the FAQ.compile file (in case you're not aware there's > > some info about compiling on SuSE in this file which should be in the root > > directory of the scigraphica source tree). I'll see about getting this > > into the docs proper over xmas. > > The webpage doesn't list imlib as a requirement. Should be either true or > changed ;) More people complained about this; indeed there was still the imlib link in the zvt/Makefile.am. I've removed that...in CVS that is! > Hmm, I don't have Gnome on my system, and as the Scigraphica webpage > suggests that it is also useable without it I didn't want to install it. > Seems a little bit strange that you have to install a sound package for a > scientific plotting system, especially on a machine which doesn't even > have a sound card... But at least I'm now able to compile the beast. configure with the "--without-gnome" option and you can do everything without the GNOME stuff! Merry Xmass, Rob. ```
 Re: [Scigraphica-devel] Solution for Bug #494545 From: Robert Schwebel - 2001-12-22 13:28:23 ```Hi John, On Sat, 22 Dec 2001, John Bland wrote: > It's true that you need the termcap link but you shouldn't have needed to > do anything for readline on 7.2. This is a known problem and is in the > FAQ.compile file. Oops, I would have expected these things in the INSTALL file and didn't look further after seeing it was just the generic one. > Hmm, gdk_imlib is supplied by the imlib-devel package, which is listed in > the requirements in the FAQ.compile file (in case you're not aware there's > some info about compiling on SuSE in this file which should be in the root > directory of the scigraphica source tree). I'll see about getting this > into the docs proper over xmas. The webpage doesn't list imlib as a requirement. Should be either true or changed ;) > The problem above isn't due to python, you don't have the db-devel package > installed. Would be great to test the availability of this stuff from autoconf. Just in case somebody has time ;-) > If you follow the instructions at the bottom of FAQ.compile and > install the dependencies listed there scigraphica should compile > without any problems on SuSE 7.2, but if you still experience problems > there are rpms available. Hmm, only if there is no other way ;) I think I have now installed everything from the list (db-devel was missing), but it still doesn't link out of the box: ----------8<---------- /usr/lib/python2.0//config/libpython2.0.a(posixmodule.o): In function `posix_tmpnam': posixmodule.o(.text+0x30f6): the use of `tmpnam_r' is dangerous, better use `mkstemp' /usr/lib/python2.0//config/libpython2.0.a(posixmodule.o): In function `posix_tempnam': posixmodule.o(.text+0x3012): the use of `tempnam' is dangerous, better use `mkstemp' /usr/lib/python2.0//config/libpython2.0.a(bsddbmodule.o): In function `newdbhashobject': bsddbmodule.o(.text+0x76): undefined reference to `__db185_open' /usr/lib/python2.0//config/libpython2.0.a(bsddbmodule.o): In function `newdbbtobject': bsddbmodule.o(.text+0x1a3): undefined reference to `__db185_open' /usr/lib/python2.0//config/libpython2.0.a(bsddbmodule.o): In function `newdbrnobject': bsddbmodule.o(.text+0x2cd): undefined reference to `__db185_open' collect2: ld returned 1 exit status make[3]: *** [scigraphica] Error 1 make[3]: Leaving directory `/home/local/scigraphica-0.8.0/src' make[2]: *** [all-recursive] Error 1 make[2]: Leaving directory `/home/local/scigraphica-0.8.0/src' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/home/local/scigraphica-0.8.0' make: *** [all-recursive-am] Error 2 ---------->8---------- __db185_open is in /usr/lib/libdb.so: callisto: # nm -n /usr/lib/libdb.so |grep __db185_open 0000f750 T __db185_open But the order of the linker argument seems to be wrong: -lpython2.0 comes before -ldb. I've temporarily added '-L/usr/lib -ldb' before '-lpython2.0' and then it compiles. The warnings are still there, but at least it works. > The configure script should guide you as to what to install if > necessary and apart from the termcap link it should compile with the > basic SuSE packages + Numeric +gtkextra without any tinkering, so just > pereservere with yast! ;0) AAARGH! Did you say "yast"? Don't use the bad word... :-) > [One thing to look out for is esound, for some reason not having > esound-devel prevents the configure script from detecting some GNOME > packages, so just install it] Hmm, I don't have Gnome on my system, and as the Scigraphica webpage suggests that it is also useable without it I didn't want to install it. Seems a little bit strange that you have to install a sound package for a scientific plotting system, especially on a machine which doesn't even have a sound card... But at least I'm now able to compile the beast. Thanks for your help! Robert -- +--------------------------------------------------------+ | Dipl.-Ing. Robert Schwebel | http://www.pengutronix.de | | Pengutronix - Linux Solutions for Science and Industry | | Braunschweiger Str. 79, 31134 Hildesheim, Germany | | Phone: +49-5121-28619-0 | Fax: +49-5121-28619-4 | +--------------------------------------------------------+ ```
 Re: [Scigraphica-devel] Solution for Bug #494545 From: John Bland - 2001-12-22 11:27:33 ```On Sat, 22 Dec 2001, John Bland wrote: > If you follow the instructions at the bottom of FAQ.compile and install > the dependencies listed there scigraphica should compile without any > problems on SuSE 7.2, but if you still experience problems there are rpms > available. Hmm, to follow myself up the instructions are for 7.1 not 7.2. There's not much difference, if you experience problems then just try installing any -devel packages that seem appropriate (readline-devel isn't listed). There are newer versions of Numeric than that listed and you don't need Distutils anymore I think. The configure script should guide you as to what to install if necessary and apart from the termcap link it should compile with the basic SuSE packages + Numeric +gtkextra without any tinkering, so just pereservere with yast! ;0) [One thing to look out for is esound, for some reason not having esound-devel prevents the configure script from detecting some GNOME packages, so just install it] Frink -- John Bland M.Phys (Hons) AMInstP / \ PhD Student & Sys Admin Email: j.bland at cmp.liv.ac.uk / \ Condensed Matter Group http://ringtail.cmp.liv.ac.uk/ / \ Liverpool University "Don't make me bite you in hard-to-reach places!" -- The Tick ```
 Re: [Scigraphica-devel] Solution for Bug #494545 From: John Bland - 2001-12-22 11:20:08 ```On Fri, 21 Dec 2001, Robert Schwebel wrote: > Hi, > > I just tried to install scigraphica-0.8.0 on a SuSE-7.2 based machine and > came across the same problem which was reported in the bug tracking system > with the number mentioned in the subject. > > The fact that configure doesn't find libtermcap (and libreadline as well) > comes from SuSE not making symbolic links from libtermcap.so.2.0.8 to > libtermcap.so in /usr/lib and from libreadline.so.4.2 to libreadline.so in > /lib. Making the links by hand repairs this. It's true that you need the termcap link but you shouldn't have needed to do anything for readline on 7.2. This is a known problem and is in the FAQ.compile file. > Next problem is that when you don't have gdk_imlib installed (which is not > in the list of required packages) you'll run into problems with compiling > the integrated zvt, as in Makefile.am there is a hardcoded -lgdk_imlib > which is not influenced by the --without-imlib switch of the toplevel > configure. I removed it by hand and it obviously works. Hmm, gdk_imlib is supplied by the imlib-devel package, which is listed in the requirements in the FAQ.compile file (in case you're not aware there's some info about compiling on SuSE in this file which should be in the root directory of the scigraphica source tree). I'll see about getting this into the docs proper over xmas. > bsddbmodule.o(.text+0x76): undefined reference to `__db185_open' > /usr/lib/python2.0//config/libpython2.0.a(bsddbmodule.o): In function > `newdbbtobject': > bsddbmodule.o(.text+0x1a3): undefined reference to `__db185_open' > /usr/lib/python2.0//config/libpython2.0.a(bsddbmodule.o): In function > `newdbrnobject': > bsddbmodule.o(.text+0x2cd): undefined reference to `__db185_open' > collect2: ld returned 1 exit status > > Can you give me a hint which version of python is required for > scigraphica? Would also be good put that on the web page... SuSE installs > a package named python-2.0-90.rpm. The problem above isn't due to python, you don't have the db-devel package installed. The versions of python with SuSE 7.2 and 7.3 seem fine IME. If you follow the instructions at the bottom of FAQ.compile and install the dependencies listed there scigraphica should compile without any problems on SuSE 7.2, but if you still experience problems there are rpms available. HTH, Frink -- John Bland M.Phys (Hons) AMInstP / \ PhD Student & Sys Admin Email: j.bland at cmp.liv.ac.uk / \ Condensed Matter Group http://ringtail.cmp.liv.ac.uk/ / \ Liverpool University "Don't make me bite you in hard-to-reach places!" -- The Tick ```
 [Scigraphica-devel] Best wishes! From: Adrian Feiguin - 2001-12-22 04:36:24 ```Dear friends, users, colleagues, developers... I am moving to Irvine, California. We are leaving tomorrow, pulling a small trailer and crossing the country behind the wheel. We'll spend Xmas somewere on the road, like San Antonio, TX, and continue our travesy toward my new life as a pour postdoc at UCI. I hope we make it and my pour car survives. I'm going to be in touch with the 600 hs of free AOL, and my wife's laptop, but I can't promess that I will solve your problems until I settle down ;-) (However, our fabulous team is going to be in touch, so don't hesitate to post your questions). I want to wish you all a merry Xmas, happy Hannukah, and a super wonderful new year (specialy for my country, snif!) Best wishes, and happy coding! PS: I'm getting my Tux t-shirt for X-mas, yuppiee! ```
 [Scigraphica-devel] Best wishes! From: Adrian Feiguin - 2001-12-21 19:29:28 ```Dear friends, users, colleagues, developers... I am moving to Irvine, California. We are leaving tomorrow, pulling a small trailer and crossing the country behind the wheel. We'll spend Xmas somewere on the road, like San Antonio, TX, and continue our travesy toward my new life as a pour postdoc at UCI. I hope we make it and my pour car survives. I'm going to be in touch with the 600 hs of free AOL, and my wife's laptop, but I can't promess that I will solve your problems until I settle down ;-) (However, our fabulous team is going to be in touch, so don't hesitate to post your questions). I want to wish you all a merry Xmas, happy Hannukah, and a super wonderful new year (specialy for my country, snif!) Best wishes, and happy coding! PS: I'm getting my Tux t-shirt for X-mas, yuppiee! ```
 Re: [Scigraphica-devel] Re: build notes so far on Irix6.5 From: Mathew Yeates - 2001-12-21 18:49:30 ```Hi Rob > > How about adding to autogen.sh: > > # Silicon Graphics/IRIX patch; prevents loading gtkextra.m4 twice > if test `uname` = IRIX -a -e [where-ever the system gtkextra.m4 is located] \ > -a -e macros/gtkextra.m4 ; then > echo > echo -n 'IRIX patch: removing "macros/gtkextra.m4"... ' > rm -f macros/gtkextra.m4 > echo "done" > echo > fi > > Please replace the [where-ever ... located] by the default-location of your > IRIX gtkextra.m4. Tell me if this works and I'll add it to CVS. > There has to be a better way than this. You are going to need to have a line for a every possible machine. What is really important is whether gtkextra.m4 is already installed. An its location can be wherever aclocal stores its macros and this can be ANYWHERE! It depends on how aclocal was configured. aclocal --print-ac-dir will tell you where. Then you can just see if gtkextra.m4 is in that dir > > 2) if either --with-termcap-path or --with_readline_path are not set, > > configure tries to compile its test program with the flag -L. On Irix, > > this causes the library search path to get deleted. > > The responsible code *could* be in configure.in line 35: > > LIBS="\$LIBS -L\$termcap_path -L\$readline_path" > > When not specified, both \$termcap_path and \$readline_path are empty, > so we could replace this by: > > if test -n "\$termcap_path" ; then > LIBS="\$LIBS -L\$termcap_path" > fi > > And a similar if-test for "\$readline_path" > > Please try and inform me if that solves the problem > (don't forget to rerun ./autgen.sh and ./configure after making these changes!) That looks like a better idea. > > > 3) need readline to be installed. I don't remember this being mentioned > > as a requirement. > > Good point! Conrad, should this dependency go somewhere in the SG info pages? > > > 4) arrayobject.h only searched for in python system path. Should be > > configurable for cases where Numeric is installed elsewhere > > I don't know whether this is linked to python's inflexibility of searching > elsewhere, or if we should add more flags to the configure script. I have no > idea! > Conrad? The problem is that configure does a "find" in the python system directory. There should be a switch to configure allowing the path to be specified. > > > 5) During config, if imlib is not found, configure says image handling > > is disabled. But when zvt is built, gdk_imlib is still linked in > > The imlib inclusions where already commented out in the zvt/zvtterm.[ch], but > the Makefile.am still added the imlib link. I've removed that also as of today. > > > > 6) msgfmt must be installed > > Yes, I believe so. We've just added the internationlization and still figuring > out > what is needed and what is not for getting this up and running. It's seems to be > working now, but I still have not a clear picture of what the real dependencies > are. > You probably need the GNU gettext package installed in one or another way. > > > > 7) in src/python includedir is passed to CFLAGS, picks up old gtkextra include files > > Don't understand this. Could you be a little more specific what is the actual > problem here? Sure. I have an old version of gtkextra installed and there are header files in ~mathew/include. My new version of gtkextra lives in its own directory ~/gtkx. I was trying to install sg into my home directory with ./configure --prefix=/hsgi/mathew but this caused the variable includedir to be set to /hsgi/mathew/include. Then - \$includedir is passed on the compile line like CC -I\$includedir and this causes the wrong headers to be included. > > > 8) no mpc library. Had to remove -lmpc. According to the Irix man > > pages there are a few parallelizing libraries but they are called > > libc_mp.a, /usr/lib64/mips3/libmp.a and /usr/lib64/mips4/libmp.a > > I have no idea what you're talking about. Please clarify. I have no libmpc. On the link line, this library appears. > > > 9) doesnt appear the readline include path is making to the makefile. > > In fact, the path isnt mentioned in the Makefile for src/python > > Sorry, what is your point here? you can set the readline include path to anything you want as a switch to configure. But this value is ignored. > > > 10) Unresolved text symbols in src directory. On the link line, > > Libraries are linked in before they > > are needed. For example, the line -L../zvt -lsgvt appears AFTER libgtk is > > linked in. > > This could be linked to the order of how the scripts are called in configure.in. > I'm not sure though. If so, you need to know which order is wrong and swap the > test script in configure.in. Solution: In the Makefile for src, add -L../zvt -lsgvt near the front of the list where LIBS is set. Also, -lz needs to be added > > -------------------------------------------------------------------------- > > Let's start with this. > Please apply the suggestions above, and remind me of the issues that > remain. We'll continue then tackling them one by one :). > > We advertise SciGraphica to compile on just any Unix-like OS; so your > help on IRIX is very welcome! > > Thanks, > Rob. > > _______________________________________________ > Scigraphica-devel mailing list > Scigraphica-devel@... > https://lists.sourceforge.net/lists/listinfo/scigraphica-devel ```
 [Scigraphica-devel] Solution for Bug #494545 From: Robert Schwebel - 2001-12-21 14:49:13 ```Hi, I just tried to install scigraphica-0.8.0 on a SuSE-7.2 based machine and came across the same problem which was reported in the bug tracking system with the number mentioned in the subject. The fact that configure doesn't find libtermcap (and libreadline as well) comes from SuSE not making symbolic links from libtermcap.so.2.0.8 to libtermcap.so in /usr/lib and from libreadline.so.4.2 to libreadline.so in /lib. Making the links by hand repairs this. Next problem is that when you don't have gdk_imlib installed (which is not in the list of required packages) you'll run into problems with compiling the integrated zvt, as in Makefile.am there is a hardcoded -lgdk_imlib which is not influenced by the --without-imlib switch of the toplevel configure. I removed it by hand and it obviously works. Now, there seem to be issues with the python-2.0 version which came with the distro: ----------8<---------- gcc -g -O2 -I -I -DREADLINE_4 -I/usr/local/include/gtk-1.2 -I/usr/local/include/glib-1.2 -I/usr/local/lib/glib/include -I/usr/X11R6/include -I/usr/local/include/gtk-1.2 -I/usr/local/include/glib-1.2 -I/usr/local/lib/glib/include -I/usr/X11R6/include -I../zvt -I../../zvt -o scigraphica gtkpixmapmenuitem.o gtkplotart.o gtkplotgnome.o sg.o sg_arrange_dialog.o sg_axis_dialog.o sg_clipboard.o sg_column_dialog.o sg_config.o sg_convert_dialog.o sg_dataset.o sg_dataset_dialog.o sg_dialogs.o sg_edit_function_dialog.o sg_edit_data_dialog.o sg_edit_columns_dialog.o sg_edit_exp_dialog.o sg_edit_3d_dialog.o sg_ellipse_dialog.o sg_entry.o sg_file.o sg_file_dialog.o sg_formula_dialog.o sg_frame_dialog.o sg_function_dialog.o sg_grids_dialog.o sg_import_dialog.o sg_labels_dialog.o sg_layer.o sg_layer_control.o sg_layer_dialog.o sg_legends_dialog.o sg_line_dialog.o sg_logo_dialogs.o sg_matrix_menu.o sg_matrix_convert.o sg_matrix_dialog.o sg_menu.o sg_misc.o sg_misc_dialogs.o sg_new_data_dialog.o sg_page_dialog.o sg_planes_dialog.o sg_plot.o sg_plot_file.o sg_plot_file_xml.o sg_plot_menu.o sg_plot_tools.o sg_plugin.o sg_preferences_dialog.o sg_project.o sg_project_menu.o sg_project_file_xml.o sg_project_file_sax.o sg_project_rescue.o sg_project_autosave.o sg_rectangle_dialog.o sg_stock.o sg_style_dialog.o sg_text_dialog.o sg_title_dialog.o sg_toggle_combos.o sg_toolbox.o sg_worksheet.o sg_worksheet_file.o sg_worksheet_file_ascii.o sg_worksheet_file_html.o sg_worksheet_file_tex.o sg_worksheet_file_xml.o sg_worksheet_menu.o sg_worksheet_tools.o sg_wrap.o sg_xy_formula_dialog.o -rdynamic -rdynamic -L/usr/local/scigraphica-0.8.0/src/python python/libpint.a -L/usr/local/lib -L/usr/X11R6/lib -ldl -lXext -lX11 -lm -ldl -lXext -lX11 -lm -ldl -ldl -lXext -lX11 -lm -ldl -lXext -lX11 -lm -ldl -lXext -lX11 -lm -ldl -lXext -lX11 -lm -lm /usr/local/lib/libxml.so -lz -L/usr/lib/python2.0//config -lpython2.0 -ldb -lpthread -ldl -lutil -lutil -L/usr/local/scigraphica-0.8.0/src -lreadline -ltermcap -lncurses -ldl -lXext -lX11 -lm -ldl -lXext -lX11 -lm /usr/local/lib/libgmodule.so -ldl -ldl -lXext -lX11 -lm -ldl -lXext -lX11 -lm -ldl -lXext -lX11 -lm /usr/local/lib/libgtkextra.so /usr/local/lib/libgtk.so /usr/local/lib/libgdk.so /usr/lib/libgmodule.so /usr/lib/libglib.so -ldl -lXext -lX11 -lm /usr/local/lib/libglib.so -lm -L/usr/local/scigraphica-0.8.0/zvt -lsgvt -Wl,--rpath -Wl,/usr/local/lib -Wl,--rpath -Wl,/usr/local/lib /usr/lib/python2.0//config/libpython2.0.a(posixmodule.o): In function `posix_tmpnam': posixmodule.o(.text+0x30f6): the use of `tmpnam_r' is dangerous, better use `mkstemp' /usr/lib/python2.0//config/libpython2.0.a(posixmodule.o): In function `posix_tempnam': posixmodule.o(.text+0x3012): the use of `tempnam' is dangerous, better use `mkstemp' /usr/lib/python2.0//config/libpython2.0.a(bsddbmodule.o): In function `newdbhashobject': bsddbmodule.o(.text+0x76): undefined reference to `__db185_open' /usr/lib/python2.0//config/libpython2.0.a(bsddbmodule.o): In function `newdbbtobject': bsddbmodule.o(.text+0x1a3): undefined reference to `__db185_open' /usr/lib/python2.0//config/libpython2.0.a(bsddbmodule.o): In function `newdbrnobject': bsddbmodule.o(.text+0x2cd): undefined reference to `__db185_open' collect2: ld returned 1 exit status make[3]: *** [scigraphica] Error 1 make[3]: Leaving directory `/home/local/scigraphica-0.8.0/src' make[2]: *** [all-recursive] Error 1 make[2]: Leaving directory `/home/local/scigraphica-0.8.0/src' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/home/local/scigraphica-0.8.0' make: *** [all-recursive-am] Error 2 callisto:/usr/local/scigraphica-0.8.0 # ---------->8---------- Can you give me a hint which version of python is required for scigraphica? Would also be good put that on the web page... SuSE installs a package named python-2.0-90.rpm. Cheers, Robert -- +--------------------------------------------------------+ | Dipl.-Ing. Robert Schwebel | http://www.pengutronix.de | | Pengutronix - Linux Solutions for Science and Industry | | Braunschweiger Str. 79, 31134 Hildesheim, Germany | | Phone: +49-5121-28619-0 | Fax: +49-5121-28619-4 | +--------------------------------------------------------+ ```
 [Scigraphica-devel] Re: build notes so far on Irix6.5 From: R. Lahaye - 2001-12-21 06:10:32 ```Mathew Yeates wrote: > > I found out how to cause it and how not to cause it. Hi Mathew, At least I myself am getting a bit overloaded with your excellent report story on installing SciGraphica on SG-Irix. So lets try to go for one thing at a time. I hope Adrian or Conrad will have some idea on that core-dump or yours with the sample loading. You've sent a list with totally 12 (?) items of compile trouble on Irix-6.5. I would be happy to discuss afew of them first: > 1) autgen,sh fails unless macros/gtkextra is removed. How about adding to autogen.sh: # Silicon Graphics/IRIX patch; prevents loading gtkextra.m4 twice if test `uname` = IRIX -a -e [where-ever the system gtkextra.m4 is located] \ -a -e macros/gtkextra.m4 ; then echo echo -n 'IRIX patch: removing "macros/gtkextra.m4"... ' rm -f macros/gtkextra.m4 echo "done" echo fi Please replace the [where-ever ... located] by the default-location of your IRIX gtkextra.m4. Tell me if this works and I'll add it to CVS. > 2) if either --with-termcap-path or --with_readline_path are not set, > configure tries to compile its test program with the flag -L. On Irix, > this causes the library search path to get deleted. The responsible code *could* be in configure.in line 35: LIBS="\$LIBS -L\$termcap_path -L\$readline_path" When not specified, both \$termcap_path and \$readline_path are empty, so we could replace this by: if test -n "\$termcap_path" ; then LIBS="\$LIBS -L\$termcap_path" fi And a similar if-test for "\$readline_path" Please try and inform me if that solves the problem (don't forget to rerun ./autgen.sh and ./configure after making these changes!) > 3) need readline to be installed. I don't remember this being mentioned > as a requirement. Good point! Conrad, should this dependency go somewhere in the SG info pages? > 4) arrayobject.h only searched for in python system path. Should be > configurable for cases where Numeric is installed elsewhere I don't know whether this is linked to python's inflexibility of searching elsewhere, or if we should add more flags to the configure script. I have no idea! Conrad? > 5) During config, if imlib is not found, configure says image handling > is disabled. But when zvt is built, gdk_imlib is still linked in The imlib inclusions where already commented out in the zvt/zvtterm.[ch], but the Makefile.am still added the imlib link. I've removed that also as of today. > 6) msgfmt must be installed Yes, I believe so. We've just added the internationlization and still figuring out what is needed and what is not for getting this up and running. It's seems to be working now, but I still have not a clear picture of what the real dependencies are. You probably need the GNU gettext package installed in one or another way. > 7) in src/python includedir is passed to CFLAGS, picks up old gtkextra include files Don't understand this. Could you be a little more specific what is the actual problem here? > 8) no mpc library. Had to remove -lmpc. According to the Irix man > pages there are a few parallelizing libraries but they are called > libc_mp.a, /usr/lib64/mips3/libmp.a and /usr/lib64/mips4/libmp.a I have no idea what you're talking about. Please clarify. > 9) doesnt appear the readline include path is making to the makefile. > In fact, the path isnt mentioned in the Makefile for src/python Sorry, what is your point here? > 10) Unresolved text symbols in src directory. On the link line, > Libraries are linked in before they > are needed. For example, the line -L../zvt -lsgvt appears AFTER libgtk is > linked in. This could be linked to the order of how the scripts are called in configure.in. I'm not sure though. If so, you need to know which order is wrong and swap the test script in configure.in. -------------------------------------------------------------------------- Let's start with this. Please apply the suggestions above, and remind me of the issues that remain. We'll continue then tackling them one by one :). We advertise SciGraphica to compile on just any Unix-like OS; so your help on IRIX is very welcome! Thanks, Rob. ```
 [Scigraphica-devel] how to duplicate my core dump From: Mathew Yeates - 2001-12-21 02:28:12 ```I found out how to cause it and how not to cause it. 1) Start sg. The window opens and says Data01 on the worksheet page. The name "Data01" is colored. 2) Click on Help/examples/simple test. The window goes blank and a dialog box shows up saying to enter a name. After the name is entered you'll get the dump. It can be avoided by clicking on the box holding the name "Data01" and hitting return. Then the examples can be viewed. Mathew ```
 [Scigraphica-devel] debugger output from core dump From: Mathew Yeates - 2001-12-20 23:09:54 ```I finally got gdb to work after I disconnected the signals in project_rescue. (dbx still doesnt work). Hope this helps where: #0 0x100a866c in sg_worksheet_rename (worksheet=0x105da6b8, name=0x105e9d70 "aa") at sg_worksheet.c:856 #1 0x10083880 in sg_project_rename_worksheet (worksheet=0x105da6b8, name=0x105e9d70 "aa") at sg_project.c:776 #2 0x1008358c in rename_worksheet (iconlist=0x105da6b8, item=0x105e1860) at sg_project.c:718 #3 0x5ff0e57c in gtk_marshal_BOOL__POINTER (object=0x105da6b8, func=0, func_data=0x0, args=0xc218020) at gtkmarshal.c:28 #4 0x5ff55878 in gtk_handlers_run (handlers=0x105e0df0, signal=0x7fff1480, object=0x105cdf50, params=0x7fff1510, after=0) at gtksignal.c:1917 #5 0x5ff54658 in gtk_signal_real_emit (object=0x105cdf50, signal_id=88, params=0x7fff1510) at gtksignal.c:1477 #6 0x5ff514fc in gtk_signal_emit (object=0x105cdf50, signal_id=88) at gtksignal.c:552 #7 0x40daadc in deactivate_entry (iconlist=0x105cdf50) at gtkiconlist.c:652 #8 0x40dda74 in gtk_icon_list_set_active_icon (iconlist=0x105cdf50, icon=0x0) at gtkiconlist.c:1639 #9 0x40dc55c in gtk_icon_list_put (iconlist=0x105cdf50, x=10, y=4, pixmap=0x106ad8d8, mask=0x10642570, label=0x105de960 "Data01", data=0x105da6b8) at gtkiconlist.c:1152 #10 0x40dc448 in gtk_icon_list_real_add (iconlist=0x105cdf50, pixmap=0x106ad8d8, mask=0x10642570, label=0x105de960 "Data01", data=0x105da6b8) at gtkiconlist.c:1126 #11 0x40dc28c in gtk_icon_list_add_from_data (iconlist=0x105cdf50, data=0x104a6348, label=0x105de960 "Data01", link=0x105da6b8) at gtkiconlist.c:1058 #12 0x1008181c in sg_project_new_worksheet () at sg_project.c:163 #13 0x10087498 in sgStartElement (state=0x7fff1f60, real_name=0x1052b1a8 "sgw:Worksheet", attrs=0x105e1860) at sg_project_file_sax.c:383 #14 0x41a1e1c in xmlParseStartTag (ctxt=0x1063c3b0) at parser.c:5969 #15 0x41a38c0 in xmlParseElement (ctxt=0x1063c3b0) at parser.c:6299 #16 0x41a3674 in xmlParseContent (ctxt=0x1063c3b0) at parser.c:6219 #17 0x41a3be0 in xmlParseElement (ctxt=0x1063c3b0) at parser.c:6369 #18 0x41a9d90 in xmlParseDocument (ctxt=0x1063c3b0) at parser.c:6984 #19 0x41ad0a8 in xmlSAXUserParseFile (sax=0x104b19a8, user_data=0x7fff1f60, filename=0x105e1860 "") at parser.c:8530 #20 0x1008e348 in sg_project_file_import_xml ( file=0x7fff2180 "/hsgi/mathew/sg/share/scigraphica/examples/simpletest.sg", worksheet=0x15, plot=0x105e1860) at sg_project_file_sax.c:2340 #21 0x100864c0 in sg_menu_SGexample (widget=0x1034e0a0, data=0x1034def0) at sg_project_menu.c:607 #22 0x5ff0ea80 in gtk_marshal_NONE__NONE (object=0x105da6b8, func=0xffffffff, func_data=0x0, args=0x2) at gtkmarshal.c:312 #23 0x5ff55878 in gtk_handlers_run (handlers=0x10547b40, signal=0x7fff2320, object=0x105c2918, params=0x7fff23b0, after=0) at gtksignal.c:1917 etc. the offending line causing the SEGV is 856 if(worksheet->icon->entry_label){ at the this point, worksheet->icon is NULL ```
 [Scigraphica-devel] still segfaulting From: Mathew Yeates - 2001-12-20 21:41:42 ```Checked out the latest from cvs, I still get Gtk-WARNING **: invalid cast from (NULL) pointer to `GtkEntry' Gtk-CRITICAL **: file gtkentry.c: line 492 (gtk_entry_get_text): assertion `entry != NULL' failed. Segmentation fault (core dumped) when I try and run the examples ```
 [Scigraphica-devel] Re: suspicious python terminal code From: Conrad Steenberg - 2001-12-20 17:54:29 ```On Thu, 2001-12-20 at 07:30, R. Lahaye wrote: > > Conrad, > > In python_term.c:create_child() contains a suspicious > use of variable "int i": > > for (;;) > { gchar buffer[256]; > int len; > do { int i = 0; > len = read(com_pipe[0],buffer,256); > if (i>10) { > write(line_pipe[1], "\n", 1); > usleep(100000); > i=0; > } > } while (len <= 0); > > (lines 615-627 in new(!) CVS version) > "i" is initialized to zero, nowhere used or touched, but > the if statement tests whether it's larger than 10 !?!? > Consequently, the if-body is NEVER executed. Is this > on purpose, or is something fishy here? > > Under present circumstaces we can remove the if-statement > together with integer i! OK, sure. > > Regards, > Rob. ```
 Re: [Scigraphica-devel] "--plot=" and "--worksheet=" args in sg.c.What for? From: Adrian Feiguin - 2001-12-20 16:16:49 ```The idea works like this: you can use sg to edit a plot or a worksheet alone, without needing the entire interface. If you have an .sgp or .spw you type scigraphica --plot myplot.sgp and sg opens only a plot window. That's what I use gui_iconlist/plot/worksheet for, although a better idea would be to use a gui_mode variable with values enum { GUI_ICONLIST, GUI_PLOT, GUIL_WORKSHEET }. The idea is also to have an interface that can be embedded in other applications. Suppouse (it's still a dream) that you have a gnumeric file and you want to plot the data using scigraphica. Then you can embedd a plot window in gnumeric, without the iconlist. This will require more work in the Bonobo interface. Saludos, On Wed, 19 Dec 2001, R. Lahaye wrote: > Adrian Feiguin wrote: > > > > On Tue, 18 Dec 2001, R. Lahaye wrote: > > > > > > Or is that --plot/--worksheet in sg.c:init_global_vars() unrelated to this? > > > > > > > This was an experiment, it worked with worksheets and plots, and now it > > has to be improved. I haven't had time though. > > Would you mind me modifying the argument list options? > > Why would we need the "--worksheet" or "--plot"? Instead let the > file-extension (".sgw" or ".sgp") determine whether we are going to > import a worksheet or a plot. Along these lines we then can also open > a project from the command line, when the file-extension is ".sg". > > Regards, > Rob. > > _______________________________________________ > Scigraphica-devel mailing list > Scigraphica-devel@... > https://lists.sourceforge.net/lists/listinfo/scigraphica-devel > ```
 [Scigraphica-devel] suspicious python terminal code From: R. Lahaye - 2001-12-20 15:29:29 ```Conrad, In python_term.c:create_child() contains a suspicious use of variable "int i": for (;;) { gchar buffer[256]; int len; do { int i = 0; len = read(com_pipe[0],buffer,256); if (i>10) { write(line_pipe[1], "\n", 1); usleep(100000); i=0; } } while (len <= 0); (lines 615-627 in new(!) CVS version) "i" is initialized to zero, nowhere used or touched, but the if statement tests whether it's larger than 10 !?!? Consequently, the if-body is NEVER executed. Is this on purpose, or is something fishy here? Under present circumstaces we can remove the if-statement together with integer i! Regards, Rob. ```
 [Scigraphica-devel] EPS size again From: Daniel Connelly - 2001-12-20 14:53:11 ```A quick experiment, on a grey-scale 2-D contour plot. The file sg.eps was produced by SciGraphica. The file gs.eps is produced by gs -- same plot. % gs -sOutputFile=gs.eps -sDEVICE=epswrite sg.eps % ls -l junk.eps n_vg0.eps -rw-r--r-- 1 djconnel djconnel 297853 Dec 20 06:47 gs.eps -rw-r--r-- 1 djconnel djconnel 2116038 Dec 20 06:45 sg.eps Previously, I developed a patch to an older version of gtk-extra (0.99.16) which allowed for automatic piping of postscript output through gs. This also allowed "direct" generation of PNG and PDF through the associated gs drivers. Sometime in the next few weeks (no time at the moment), I'll try to port that to the new gtk-extra. Obviously, for publication, the resulting PS compression is quite significant, since there can be a lot of figures in one paper. Dan ```
 [Scigraphica-devel] Tutorial 3 From: John Bland - 2001-12-20 13:26:57 ```Tutorial 3 has been added to CVS in SGML format, along with some necessary figures (not many). Something I came across during this was that binary files should be added with the -kb switch in case cvs gets confused reading them, just in case anyone else wanted to put binary files in the repository. I'll be offline from about 22nd Dec to ~5th Jan but will try to update a few of the docs (if I'm not so ratfaced I can't even type ;0). Merry xmas and all that. Frink -- John Bland M.Phys (Hons) AMInstP / \ PhD Student & Sys Admin Email: j.bland at cmp.liv.ac.uk / \ Condensed Matter Group http://ringtail.cmp.liv.ac.uk/ / \ Liverpool University "Don't make me bite you in hard-to-reach places!" -- The Tick ```
 [Scigraphica-devel] Re: Another Ctrl-D bug fix attempt From: R. Lahaye - 2001-12-20 02:12:22 ```Conrad Steenberg wrote: > > On Wed, 2001-12-19 at 14:49, R. Lahaye wrote: > > Could you give me a quick hint as to where and how you prefer to have > > the quit command interpreted? > > All the python commands are in src/python/python_sheet.c. Just let the > command call src/python/python_term.c:term_kill_child(). Sorry I'm lost in the python code now: How do these functions in python_sheet.c connect to what is actually typed in the python-terminal? Simply adding a function "quit(...)" in python_sheet.c is apparently not sufficient. Can't figure the missing link? Thanks, Rob. ```
 [Scigraphica-devel] Re: Another Ctrl-D bug fix attempt From: Conrad Steenberg - 2001-12-20 01:15:54 ```On Wed, 2001-12-19 at 14:49, R. Lahaye wrote: > Conrad Steenberg wrote: > > > > On Wed, 2001-12-19 at 08:42, R. Lahaye wrote: > > > We can only quit the python terminal by using its emergency exit, i.e. > > > the windowmanager's destroy. There's no clean way for exiting the > > > python terminal. > > > > Why is that not clean? Anyway, that is not a bug report, it is a feature > > request ;-) > > > > Yes, this is solvable, and you're welcome to solve it if it is a > > priority for you. > > Sure, I can give it a shot. > > Could you give me a quick hint as to where and how you prefer to have > the quit command interpreted? > > (I don't think I have noticed a command interpreter in the terminal > code; it doesn't exist yet? Will this be the first command interpreted?) All the python commands are in src/python/python_sheet.c. Just let the command call src/python/python_term.c:term_kill_child(). Cheers! Conrad ```

Showing results of 190

1 2 3 .. 8 > >> (Page 1 of 8)