Re: [Ifeffit] Artemis: Error in laoding additional FEFF paths?
Matt, as to the following question: QUESTION: I do not seem to have, in tcsh, the command "configure" I think I solved this one by entering the /Users/stefano/Tk-804.027/ifeffit-1.2.5 directory I am proceeding with the resto of the indicated protocol, so far not caring about the downloading directory (I am using my home directory right now). Please let me know whether I can change the Tk-804.027 directory from my home directory to where it is more right to move it (which I do not know). Thanks, Stefano -- ____________________________________________ Stefano Ciurli Professor of Chemistry Department of Agro-Environmental Science and Technology University of Bologna Viale Giuseppe Fanin, 40 I-40127 Bologna Italy Phone: +39-051-209-6204 Fax: +39-051-209-6203 "Fatti non foste a viver come bruti, ma per seguir virtute e canoscenza" Dante Alighieri - Inferno - Canto XXVI
Hi Stefano, Sorry for the incomplete instructions and not emphasizing some of the important points of the Unix shell. I believe Paul answered most of this, and it seems like you've either got it working or are close, but I'll answer anyway:
2. install Perl/Tk from source: 2.1 download Tk-804.027.tar.gz from a CPAN site ~> wget http://www.cpan.org/modules/by-authors/id/NI-S/Tk-804.027.tar.gz
QUESTION: is there a directory where I should download the file and then follow up with the subsequent instructions?
No, it doesn't matter. Using a temporary directory that you later delete would be fine: ~> mkdir MyTemp ~> cd MyTemp ~> wget http://www.cpan.org/modules/by-authors/id/NI-S/Tk-804.027.tar.gz ~> tar xvzf Tk-804.027.tar.gz ~> cd Tk-804.027 ~> ### change pTk/mTk/tixGeneric/tixHList.c as earlier instructions ~> perl Makefile.PL ; make ~> sudo make install ~> cd.. ~> wget http://cars9.uchicago.edu/ifeffit/src/ifeffit-1.2.5.tar.gz ~> tar xvzf ifeffit-1.2.5.tar.gz ~> cd ifeffit-1.2.5 ~> ./configure --prefix=/Applications/Ifeffit ## NOTE leading './' !!! ~> make ~> sudo make install ~> . /Applications/Ifeffit/bin/iff_init.sh ## NOTE leading '.' for bash ~> source /Applications/Ifeffit/bin/iff_init.csh ~> horae_update --force ~> cd .. At this point, you can delete the MyTemp whole folder it. Or you could keep the *.tar.gz files in case anything goes wrong (never!) and just delete the Tk-804.027 and ifeffit-1.2.5 folders.
another QUESTION: are there any files I should delete and from which directory before doing the whole thing from scratch?
No, everything should be overwritten as needed.
Aquaterm and 1.2.5??
With ifeffit-1.2.5a for Mac OS X, the Aquaterm plotting should work 'out of the box'. The Aquaterm device is now the default device as it has several advantages over the X11 device.
Also: Is the question related to which Tk/perl version I should have installed still valid? It is not clear, from your web page, whether the comments for the Linux installation are also related to the OSX.
To be honest, I'm not sure. I think we might have to wait for Bruce on this one. As far as I understand it, some of the artemis bugs were related to the Perl/Tk versions and to the tixHList bug, but I'm not sure all of them were. Hope that helps, --Matt
Aquaterm and 1.2.5??
With ifeffit-1.2.5a for Mac OS X, the Aquaterm plotting should work 'out of the box'. The Aquaterm device is now the default device as it has several advantages over the X11 device.
Hi Matt, Does Aquaterm now respond to the cursor command? Jeff
Hi folks, I now have a running set of programs, for which I thank the people who responded to my calls for help. The only anomalous behavior concerns two items: 1) After an initial run in which aquaterm started smoothly by plotting data from athena or artemis (i do not remember which one), right now for unknown reasons I need to start aquaterm before it receives orders from athena or artemis to plot data. That is, aquaterm needs to be opened first for things to go smoothly. And I do not know why 2) I do not have the ZnO tutorial files. Any one has ideas about why? Have they been removed from the package because they were giving problems? Ciao, Stefano -- ____________________________________________ Stefano Ciurli Professor of Chemistry Department of Agro-Environmental Science and Technology University of Bologna Viale Giuseppe Fanin, 40 I-40127 Bologna Italy Phone: +39-051-209-6204 Fax: +39-051-209-6203 "Fatti non foste a viver come bruti, ma per seguir virtute e canoscenza" Dante Alighieri - Inferno - Canto XXVI
I was trying to compile ifeffit 1.2.5a again on top of Matt's installer based pgplot and I am having some trouble. I have fink installed and hence another pgplot. In order not to use my fink pgplot, it is necessary to use the configure options: ./configure --prefix=/Applications/Ifeffit --with-pgplot=/Applications/Ifeffit/lib --with-pgplot-link=/Applications/Ifeffit/lib/libpgplot_iff.a with this set of options the compile proceeds smoothly until it reaches the src/test directory upon which there are a lot of linking errors (shown below). Any advice Matt? cc -g -O2 -o ifftest1 ifftest1.o ../lib/libifeffit.a /Applications/Ifeffit/lib/libpgplot_iff.a -L/sw/lib/gcc/powerpc-apple-darwin7.3.0/3.4.0 -L/sw/lib/gcc/powerpc-apple-darwin7.3.0/3.4.0/../../.. -lfrtbegin -lg2c -lSystem ld: Undefined symbols: ..objc_class_name_AQTAdapter ..objc_class_name_NSAutoreleasePool ..objc_class_name_NSConstantString ..objc_class_name_NSString _AQTRoundLineCapStyle _NSLog _NSPointFromString _NSStringFromRect _NSStringFromSize _NSZeroPoint __NSConstantStringClassReference _objc_msgSend restFP saveFP _png_create_info_struct _png_create_write_struct _png_destroy_write_struct _png_init_io _png_set_IHDR _png_set_PLTE _png_set_tRNS _png_set_text _png_write_end _png_write_info _png_write_row _XAllocColor _XChangeGC _XCheckWindowEvent _XClearWindow _XCloseDisplay _XCopyArea _XCreateGC _XCreateImage _XCreateSimpleWindow _XDefineCursor _XDeleteProperty _XDisplayName _XDrawLine _XDrawPoint _XDrawRectangle _XFillArc _XFillPolygon _XFillRectangle _XFlush _XFree _XFreeColors _XFreeGC _XGetErrorText _XGetSelectionOwner _XGetVisualInfo _XGetWindowAttributes _XGetWindowProperty _XInternAtom _XLookupString _XMapRaised _XNextEvent _XOpenDisplay _XPutImage _XQueryPointer _XResizeWindow _XSelectInput _XSendEvent _XSetErrorHandler _XSetForeground _XSetWindowBackground _XStoreColors _XSync _XVisualIDFromVisual _XWarpPointer make: *** [ifftest1] Error 1
Hi Paul, Oh, of course. My mistake. The fix is to replace ifeffit-1.2.5/iconf_pgplot with the version below. The distributed version of ifeffit-1.2.5.tar.gz is from February, and its iconf_pgplot looks under /sw for pgplot. The newer one (from April) looks under /Applications/Ifeffit. You can test it by simply typing (in the ifeffit-1.2.5 directory): ~>sh iconf_pgplot which should report something very close to -L/Applications/Ifeffit/pgplot -L/Applications/Ifeffit/lib -lpgplot_iff -lpng_iff -lz_iff -laquaterm -L/sw/lib -lg2c -L/usr/lib -lgcc -L/usr/X11R6/lib -lX11 -laquaterm -Wl,-framework -Wl,Foundation At that point, './configure; make' ought to work. Of course -- how could the Feb source kit possibly know about the Aquaterm changes made in April?? This means the way I described building from source to Stefano probably cannot work, and may be related to some of his troubles -- but how did he get it to compile is a bit of a mystery to me. I'm confused by this, as the instructions I gave Stefano should have not worked. For the record, the binary installer should work fine too. Is there evidence that it doesn't? Sorry, --Matt iconf_pgplot: #!/bin/sh # # This file will test how to build pgplot. Essentially, # this looks for where libpgplot is, where the X libs are, # and whether to use '-lpng -lz' for the PNG device. # # first, the PGPLOT directory and PGPLOT library are found # look for library file # where to look for pgplot directory: last one found will be used! pgdirs="/usr/lib /usr/local/lib /usr/share/pgplot /usr/local/pgplot /Applications/Ifeffit/pgplot $1" xdirs="/usr/X11/lib /usr/lib/X11 /usr/lib/X11R5 /usr/X11R5/lib /usr/lib/X11R6 /usr/X11R6/lib" xd="" # set default, in case we find nothing: cwd=`pwd` def="$cwd/src/pgstub/libnopgplot.a" basedir="NONE" for d in $pgdirs; do if [ -d $d ] ; then if [ -f $d/grfont.dat ] ; then basedir=$d ; fi if [ -f $d/libpgplot_iff.a ] ; then libdir=$d ; fi if [ -f $d/libpgplot.a ] ; then libdir=$d ; fi fi done # if we can't find the basedir, give up if test "$basedir" = NONE ; then echo $def ; exit ; fi # now we look for the files makefile and grexec.f # in the pgplot directory, and parse them to figure out how to # link to PGPLOT. The makefile is much more reliable, but we # can probably guess from grexec.f out="" mfile=$basedir/makefile ffile=$basedir/grexec.f if [ -f $mfile ] ; then out="$out -L$libdir" l1=`grep "^PGPLOT_LIB=" $mfile| sed 's/PGPLOT_LIB=-L\`pwd\`//' | sed 's/PGPLOT_LIB=//'` l2=`grep "^LIBS=" $mfile | sed "s/LIBS=//"` lline="$l1 $l2" for i in $lline ; do echo $i | grep -q '^-L' ; stat=$? if [ $stat == 0 ] ; then # this is a '-L' word, check that directory exists if [ -d `echo $i | sed "s/-L//"` ] ; then out="$out $i" ; fi else out="$out $i" fi done else if [ -f $ffile ] ; then out="-L$libdir -lpgplot" grep " CALL PNDRIV" $ffile >/dev/null 2>&1; if [ $? ] ; then out="$out -lpng -lz" ; fi grep " CALL XWDRIV" $ffile >/dev/null 2>&1; if [ $? ] ; then for d in $xdirs; do if [ -d $d ] ; then if [ -f $d/libX11.so ] ; then xd=$d ; fi if [ -f $d/libX11.a ] ; then xd=$d ; fi fi done out="$out -L$xd -lX11" fi else out=$def fi fi echo $out ##
Hello folks, I think I have to respond to some of the questions regarding the installation protocol I followed to install the programs on my G4: The protocol is the following: 0. install fink and fink's g77 and wget. COMMENT: I have a fink up-to-date and installed wget using "fink install wget" 1. install the Ifeffit binary from ifeffit-1.2.5a_Mac10.3.dmg. This step is to install PGPLOT correctly: the problem in question is not with PGPLOT. COMMENT: I downloaded the ifeffit-1.2.5a installer and ran the packaged installer; no problem. 2. install Perl/Tk from source: 2.1 download Tk-804.027.tar.gz from a CPAN site ~> wget http://www.cpan.org/modules/by-authors/id/NI-S/Tk-804.027.tar.gz 2.2 unpack Tk-804.027 ~> tar xvzf Tk-804.027.tar.gz 2.3 change line 2809 of pTk/mTk/tixGeneric/tixHList.c from '#if 0' to '#if 1' 2.4 build Perl/Tk: ~>perl Makefile.PL ; make 2.5 install Perl/Tk ~>sudo make install COMMENT: I followed exactly Matt's instructions as follows: ~> mkdir MyTemp ~> cd MyTemp ~> wget http://www.cpan.org/modules/by-authors/id/NI-S/Tk-804.027.tar.gz ~> tar xvzf Tk-804.027.tar.gz ~> cd Tk-804.027 ~> ### change pTk/mTk/tixGeneric/tixHList.c as earlier instructions ~> perl Makefile.PL ; make ~> sudo make install then: 3. install ifeffit from source: 3.1 download: ~>wget http://cars9.uchicago.edu/ifeffit/src/ifeffit-1.2.5.tar.gz 3.2 unpack: ~>tar xvzf ifeffit-1.2.5.tar.gz 3.3 configure ~>./configure --prefix=/Applications/Ifeffit 3.4 build ~> make 3.5 install ~> sudo make install COMMENT: It is not clear to me the difference between 1.2.5a and 1.2.5: it seems odd to me to first install a more advanced version of ifeffit (1.2.5a) and then install the older version (1.2.5). But I did so as instructed. Any explanation to this? 4. install horae 4.1 set up to use ifeffit: ~>source /Applications/Ifeffit/bin/iff_init.csh 4.2 install horae ~>horae_update --force COMMENT: At the end of this procedure I find myself with horae-033! I have to do horae_update again to go to horae-066...
I was trying to compile ifeffit 1.2.5a again on top of Matt's installer based pgplot and I am having some trouble. I have fink installed and hence another pgplot. In order not to use my fink pgplot, it is necessary to use the configure options: ./configure --prefix=/Applications/Ifeffit --with-pgplot=/Applications/Ifeffit/lib --with-pgplot-link=/Applications/Ifeffit/lib/libpgplot_iff.a
COMMENT to Paul: I guess I also have pgplot from fink installed (I simply ran the fink installation once, updating it from time to time. Now, as said above, it is up-to-date.
with this set of options the compile proceeds smoothly until it reaches the src/test directory upon which there are a lot of linking errors (shown below). Any advice Matt?
I do not see this message. Oh my, this is the first time that something works for me and does not for others! :-)) I checked: under the /sw/Applications directory, I have only one file, named AquaTerm.app. Consider that I also have an aquaterm application under the /Applications directory. I am not sure which of these two applications the program points to. Does this tell anything to anyone to explain the strange behavior I see or the differences between my installation and others'? Matt then said to Paul: The fix is to replace ifeffit-1.2.5/iconf_pgplot with the version below. The distributed version of ifeffit-1.2.5.tar.gz is from February, and its iconf_pgplot looks under /sw for pgplot. The newer one (from April) looks under /Applications/Ifeffit. You can test it by simply typing (in the ifeffit-1.2.5 directory): ~>sh iconf_pgplot COMMENT: Under /sw I do not have any ifeffit-1.2.5 directory! So I cannot run this command. Does anyone know why? Also, I do not know what to do with the iconf_pgplot text that Matt provided.... which should report something very close to -L/Applications/Ifeffit/pgplot -L/Applications/Ifeffit/lib -lpgplot_iff -lpng_iff -lz_iff -laquaterm -L/sw/lib -lg2c -L/usr/lib -lgcc -L/usr/X11R6/lib -lX11 -laquaterm -Wl,-framework -Wl,Foundation At that point, './configure; make' ought to work. Of course -- how could the Feb source kit possibly know about the Aquaterm changes made in April?? This means the way I described building from source to Stefano probably cannot work, and may be related to some of his troubles -- but how did he get it to compile is a bit of a mystery to me. I'm confused by this, as the instructions I gave Stefano should have not worked. For the record, the binary installer should work fine too. Is there evidence that it doesn't? If you, Matt, refer to the installer found in the disk image (.dmg), yes, it runs smoothly. What it does exaclt, well, too difficult for me. Stefano -- ____________________________________________ Stefano Ciurli Professor of Chemistry Department of Agro-Environmental Science and Technology University of Bologna Viale Giuseppe Fanin, 40 I-40127 Bologna Italy Phone: +39-051-209-6204 Fax: +39-051-209-6203 "Fatti non foste a viver come bruti, ma per seguir virtute e canoscenza" Dante Alighieri - Inferno - Canto XXVI
Stefano, There's a lot of confusion here, so let's go back a few steps. First, the binary installer ifeffit-1.2.5a-Mac10.3.dmg does install a complete version of Ifeffit 1.2.5 with Perl/Tk 804.027. You asked specifically about how to build this from source. I'm not sure why you want to do this, but that's OK: building from source is a fine thing to do. But let me emphasize that the goal of doing this would be to reproduce exactly what comes in the binary installer. Second, aside from g77 for building, Ifeffit relies and should use NOTHING from Fink. (Fink's wget makes the description of downloading easier, but is not needed). I think this cannot be stressed enough so I will repeat it: At no point, should Ifeffit use PGPLOT, Aquaterm, or anything else from Fink. If it does, there was a build error. As Paul pointed out, using the ifeffit-1.2.5.tar.gz source kit should not have worked. I don't understand what happened for you, but perhaps it used PGPLOT from fink. In any event, you need to rebuild from source (the 'configure' and 'make install' steps) using the newer version of iconf_pgplot, which should replace the one in the base ifeffit-1.2.5 folder -- that is, in the source directory. Again, rebuilding this will build the ifeffit library as distributed in the binary installer. If this is confusing, perhaps you should explain why you want to build from source as opposed to using the binary installer. --Matt
Just a few other words on what I did to get artemis compiled locally. I didn't reinstall Tk (although I think I could without any trouble as there I done it many times before). I did, however, compile ifeffit 1.2.5a -- note that my ifeffit source is 1.2.5a not 1.2.5 -- I don't recall the exact timing (searching through the archives of this list probably will give the answer). I suspect that this has some aquaterm related changes to it although I am not sure of the specifics. After replacing iconf_pgplot with the version in Matt's recent posting to this list, the compile, link, and install proceeded without error and I then did a "sudo horae_update -force" which resulted in a successful build and install of artemis and artemis. I did a quick check of ZnO example file number one and it worked as far as I could tell. The horae_update I did forced a new download of horae-036 as well as a build and install. Looking in that directory, I can see that there still is an examples folder and the ZnO examples are there: pwd /Volumes/swap/Ports/horae-036 [Quark:~/Ports/horae-036] paulfons% ls -lR examples/artemis/ZnO examples/artemis/ZnO: total 572 -rw-r--r-- 1 paulfons staff 1758 Dec 9 2003 README -rw-r--r-- 1 paulfons staff 6411 Dec 9 2003 README.rtf -rw-r--r-- 1 paulfons staff 599 May 1 08:58 ZnO-atoms.inp -rw-r--r-- 1 paulfons staff 12863 Nov 11 2003 ZnO.chi -rw-r--r-- 1 paulfons staff 174875 Nov 2 2003 ZnO.prj -rw-r--r-- 1 paulfons staff 9190 Jan 24 05:26 ZnO1.apj -rw-r--r-- 1 paulfons staff 90143 Jan 24 05:15 ZnO2.apj -rw-r--r-- 1 paulfons staff 88803 Jan 24 05:16 ZnO3.apj -rw-r--r-- 1 paulfons staff 93264 Jan 24 05:17 ZnO4.apj -rw-r--r-- 1 paulfons staff 85047 Jan 24 05:18 ZnOFinal.apj On Jun 15, 2004, at 8:15 PM, Stefano Ciurli wrote:
Hello folks, I think I have to respond to some of the questions regarding the installation protocol I followed to install the programs on my G4:
The protocol is the following:
0. install fink and fink's g77 and wget.
COMMENT: I have a fink up-to-date and installed wget using "fink install wget"
1. install the Ifeffit binary from ifeffit-1.2.5a_Mac10.3.dmg. This step is to install PGPLOT correctly: the problem in question is not with PGPLOT.
COMMENT: I downloaded the ifeffit-1.2.5a installer and ran the packaged installer; no problem.
2. install Perl/Tk from source: 2.1 download Tk-804.027.tar.gz from a CPAN site ~> wget http://www.cpan.org/modules/by-authors/id/NI-S/Tk-804.027.tar.gz 2.2 unpack Tk-804.027 ~> tar xvzf Tk-804.027.tar.gz 2.3 change line 2809 of pTk/mTk/tixGeneric/tixHList.c from '#if 0' to '#if 1' 2.4 build Perl/Tk: ~>perl Makefile.PL ; make 2.5 install Perl/Tk ~>sudo make install
COMMENT: I followed exactly Matt's instructions as follows:
~> mkdir MyTemp ~> cd MyTemp ~> wget http://www.cpan.org/modules/by-authors/id/NI-S/Tk-804.027.tar.gz ~> tar xvzf Tk-804.027.tar.gz ~> cd Tk-804.027 ~> ### change pTk/mTk/tixGeneric/tixHList.c as earlier instructions ~> perl Makefile.PL ; make ~> sudo make install
then:
3. install ifeffit from source: 3.1 download: ~>wget http://cars9.uchicago.edu/ifeffit/src/ifeffit-1.2.5.tar.gz 3.2 unpack: ~>tar xvzf ifeffit-1.2.5.tar.gz 3.3 configure ~>./configure --prefix=/Applications/Ifeffit 3.4 build ~> make 3.5 install ~> sudo make install
COMMENT: It is not clear to me the difference between 1.2.5a and 1.2.5: it seems odd to me to first install a more advanced version of ifeffit (1.2.5a) and then install the older version (1.2.5). But I did so as instructed. Any explanation to this?
4. install horae 4.1 set up to use ifeffit: ~>source /Applications/Ifeffit/bin/iff_init.csh 4.2 install horae ~>horae_update --force
COMMENT: At the end of this procedure I find myself with horae-033! I have to do horae_update again to go to horae-066...
I was trying to compile ifeffit 1.2.5a again on top of Matt's installer based pgplot and I am having some trouble. I have fink installed and hence another pgplot. In order not to use my fink pgplot, it is necessary to use the configure options: ./configure --prefix=/Applications/Ifeffit --with-pgplot=/Applications/Ifeffit/lib --with-pgplot-link=/Applications/Ifeffit/lib/libpgplot_iff.a
COMMENT to Paul: I guess I also have pgplot from fink installed (I simply ran the fink installation once, updating it from time to time. Now, as said above, it is up-to-date.
with this set of options the compile proceeds smoothly until it reaches the src/test directory upon which there are a lot of linking errors (shown below). Any advice Matt?
I do not see this message. Oh my, this is the first time that something works for me and does not for others! :-))
I checked: under the /sw/Applications directory, I have only one file, named AquaTerm.app. Consider that I also have an aquaterm application under the /Applications directory. I am not sure which of these two applications the program points to.
Does this tell anything to anyone to explain the strange behavior I see or the differences between my installation and others'?
Matt then said to Paul:
The fix is to replace ifeffit-1.2.5/iconf_pgplot with the version below. The distributed version of ifeffit-1.2.5.tar.gz is from February, and its iconf_pgplot looks under /sw for pgplot. The newer one (from April) looks under /Applications/Ifeffit. You can test it by simply typing (in the ifeffit-1.2.5 directory): ~>sh iconf_pgplot
COMMENT: Under /sw I do not have any ifeffit-1.2.5 directory! So I cannot run this command. Does anyone know why? Also, I do not know what to do with the iconf_pgplot text that Matt provided....
which should report something very close to -L/Applications/Ifeffit/pgplot -L/Applications/Ifeffit/lib -lpgplot_iff -lpng_iff -lz_iff -laquaterm -L/sw/lib -lg2c -L/usr/lib -lgcc -L/usr/X11R6/lib -lX11 -laquaterm -Wl,-framework -Wl,Foundation
At that point, './configure; make' ought to work.
Of course -- how could the Feb source kit possibly know about the Aquaterm changes made in April?? This means the way I described building from source to Stefano probably cannot work, and may be related to some of his troubles -- but how did he get it to compile is a bit of a mystery to me. I'm confused by this, as the instructions I gave Stefano should have not worked. For the record, the binary installer should work fine too. Is there evidence that it doesn't?
If you, Matt, refer to the installer found in the disk image (.dmg), yes, it runs smoothly. What it does exaclt, well, too difficult for me.
Stefano --
____________________________________________
Stefano Ciurli Professor of Chemistry Department of Agro-Environmental Science and Technology University of Bologna Viale Giuseppe Fanin, 40 I-40127 Bologna Italy Phone: +39-051-209-6204 Fax: +39-051-209-6203
"Fatti non foste a viver come bruti, ma per seguir virtute e canoscenza" Dante Alighieri - Inferno - Canto XXVI _______________________________________________ Ifeffit mailing list Ifeffit@millenia.cars.aps.anl.gov http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
Dr. Paul Fons Senior Researcher National Institute for Advanced Industrial Science & Technology METI Center for Applied Near-Field Optics Research (CANFOR) AIST Central 4, Higashi 1-1-1 Tsukuba, Ibaraki JAPAN 305-8568 tel. +81-298-61-5636 fax. +81-298-61-2939 email: paul-fons@aist.go.jp The lines below are in a Japanese font 〒３０５−８５６８ 茨城県つくば市東１−１−1 つくば中央第４ 近接場光応用工学センター ポール・フォンス主任研究官
Hello folks, the reasons why I wanted to build from source instead of using the binary installer was due to the fact that I was trying to get the ZnO tutorials to run smoothly as I was told happened in Paul's and Bruce's G4. I thought that would ask for a detailed set of installation instructions. Now I understand better, so I did erase the /Applications/Ifeffit directory, deleted the /Library/perl directory and re-run the binary installer for 1.2.5a I checked and the aquaterm application started smoothly when called from athena or artemis. I had in the meantime deleted the aquaterm in fink, just in case... Afterwards I ran the horae_update --force command, as told by Paul, and now I think things should go OK. I confirm the presence of the ZnO tutorial as indicated by Paul: % pwd /Applications/Ifeffit/bin/horae-036/examples/artemis/ZnO % ls README ZnO-atoms.inp ZnO.prj ZnO2.apj ZnO4.apj README.rtf ZnO.chi ZnO1.apj ZnO3.apj ZnOFinal.apj % ls -la total 1144 drwxr-xr-x 12 500 100 408 2 Jun 15:09 . drwxr-xr-x 7 500 100 238 2 Jun 15:09 .. -rw-r--r-- 1 500 100 1758 8 Dec 2003 README -rw-r--r-- 1 500 100 6411 8 Dec 2003 README.rtf -rw-r--r-- 1 500 100 599 1 May 01:58 ZnO-atoms.inp -rw-r--r-- 1 500 100 12863 10 Nov 2003 ZnO.chi -rw-r--r-- 1 500 100 174875 1 Nov 2003 ZnO.prj -rw-r--r-- 1 500 100 9190 23 Jan 21:26 ZnO1.apj -rw-r--r-- 1 500 100 90143 23 Jan 21:15 ZnO2.apj -rw-r--r-- 1 500 100 88803 23 Jan 21:16 ZnO3.apj -rw-r--r-- 1 500 100 93264 23 Jan 21:17 ZnO4.apj -rw-r--r-- 1 500 100 85047 23 Jan 21:18 ZnOFinal.apj So now I am back to where I was a few days ago. Now I start artemis by clicking on the icon and read in the ZnO.chi data, and guess what: aquaterm does NOT start. The conclusion would be that the problem occurs when I update horae to horae-036. However: I erased again the /Applications/Ifeffit directory, deleted the /Library/perl directory and re-run the binary installer for 1.2.5a Now, the ZnO directory is no longer present in the directory above. I had saved it in another directory, but when I run artemis by clicking the icon and try to read in the files in the ZnO directiy, the aquaterm does NOT start!!! Please explain this to me. I ran the horae_update command, and again the aquaterm does not start when I try to read in spectra I ran the horae_update --force command and again the aquaterm does not start when I try to read in spectra Now, I deleted again the /Applications/Ifeffit and the /Library/perl directories and re-run the binary installer, as well as the horae_update --force command and again the aquaterm does not start when I try to read in spectra I deleted again the same two directories and re-run the binary installer: when I start athena from the icon, and I try to read in data, the aquaterm does not start when I try to read in spectra. I am now in a condition where I cannot get the aquaterm to start in any of the endless trials I did. You guys keep telling me to replace iconf_pgplot with the version in Matt's recent posting to this list. I do not have any file that looks like that... I am lost. Stefano -- ____________________________________________ Stefano Ciurli Professor of Chemistry Department of Agro-Environmental Science and Technology University of Bologna Viale Giuseppe Fanin, 40 I-40127 Bologna Italy Phone: +39-051-209-6204 Fax: +39-051-209-6203 "Fatti non foste a viver come bruti, ma per seguir virtute e canoscenza" Dante Alighieri - Inferno - Canto XXVI
Stefano,
2) I do not have the ZnO tutorial files. Any one has ideas about why? Have they been removed from the package because they were giving problems?
This is at Bruce's pages for artemis, under 'contributions'.
I now have a running set of programs, for which I thank the people who responded to my calls for help. The only anomalous behavior concerns two items:
Given Paul's message, I'm now confused. Did ifeffit-1.2.5 build for you? Related to that, what was the evidence that the binary installer failed?
1) After an initial run in which aquaterm started smoothly by plotting data from athena or artemis (i do not remember which one), right now for unknown reasons I need to start aquaterm before it receives orders from athena or artemis to plot data. That is, aquaterm needs to be opened first for things to go smoothly. And I do not know why
I don't see this behavior and have never had to start Aquaterm explicitly. I can run multiple Ifeffit apps, each getting a different plot window and can close any of the plot windows: subsequent plots will create a new plot window as needed. I cannot close the Aquaterm App while running, as subsequent plots cause crashes of the calling app (this is not unique to Ifeffit. Gnuplot also crashes when Aquaterm is closed). But this does not happen normally anyway: once started, Aquaterm stays running for me even after exiting all apps using it. I suspect you still have a build problem related to the iconf_pgplot script. For what it's worth Aquaterm is the default plotting device with ifeffit-1.2.5a-Mac10.3, but the X11 device can also be used. --Matt
If the aquaterm library support is included in the pgplot distribution that the installer "installs" then perhaps changing the environment variable to specify /aqt as the pgplot device will cause pgplot to attempt to route graphics to aquaterm (hence aquaterm did not start automatically). I am still confused why Stefano did not get a bunch of linking errors though... I must say though that I really do not understand the dynamic library business -- could it be that the system just looked for the unlinked references at runtime in the Ifeffit/lib directory (as it is on the dynamic library path?). This is something I *would* like to understand... On Jun 15, 2004, at 12:17 PM, Matt Newville wrote:
Stefano,
2) I do not have the ZnO tutorial files. Any one has ideas about why? Have they been removed from the package because they were giving problems?
This is at Bruce's pages for artemis, under 'contributions'.
I now have a running set of programs, for which I thank the people who responded to my calls for help. The only anomalous behavior concerns two items:
Given Paul's message, I'm now confused. Did ifeffit-1.2.5 build for you? Related to that, what was the evidence that the binary installer failed?
1) After an initial run in which aquaterm started smoothly by plotting data from athena or artemis (i do not remember which one), right now for unknown reasons I need to start aquaterm before it receives orders from athena or artemis to plot data. That is, aquaterm needs to be opened first for things to go smoothly. And I do not know why
I don't see this behavior and have never had to start Aquaterm explicitly. I can run multiple Ifeffit apps, each getting a different plot window and can close any of the plot windows: subsequent plots will create a new plot window as needed. I cannot close the Aquaterm App while running, as subsequent plots cause crashes of the calling app (this is not unique to Ifeffit. Gnuplot also crashes when Aquaterm is closed). But this does not happen normally anyway: once started, Aquaterm stays running for me even after exiting all apps using it.
I suspect you still have a build problem related to the iconf_pgplot script. For what it's worth Aquaterm is the default plotting device with ifeffit-1.2.5a-Mac10.3, but the X11 device can also be used.
--Matt
_______________________________________________ Ifeffit mailing list Ifeffit@millenia.cars.aps.anl.gov http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
Dr. Paul Fons Senior Researcher National Institute for Advanced Industrial Science & Technology METI Center for Applied Near-Field Optics Research (CANFOR) AIST Central 4, Higashi 1-1-1 Tsukuba, Ibaraki JAPAN 305-8568 tel. +81-298-61-5636 fax. +81-298-61-2939 email: paul-fons@aist.go.jp The lines below are in a Japanese font 〒３０５−８５６８ 茨城県つくば市東１−１−1 つくば中央第４ 近接場光応用工学センター ポール・フォンス主任研究官
Matt,
I now have a running set of programs, for which I thank the people who responded to my calls for help. The only anomalous behavior concerns two items:
Given Paul's message, I'm now confused. Did ifeffit-1.2.5 build for you? Related to that, what was the evidence that the binary installer failed?
See my previous message. I did not encounter any problem following your instructions. Regarding the aquaterm issue, see my previous message concerning the two positions in the directories where I have aquaterm placed.
I don't see this behavior and have never had to start Aquaterm explicitly.
neither had I, until the latest set of installation instructions were given.
I can run multiple Ifeffit apps, each getting a different plot window and can close any of the plot windows: subsequent plots will create a new plot window as needed. I cannot close the Aquaterm App while running, as subsequent plots cause crashes of the calling app (this is not unique to Ifeffit. Gnuplot also crashes when Aquaterm is closed). But this does not happen normally anyway: once started, Aquaterm stays running for me even after exiting all apps using it.
Exactly the same behavior that I have seen until the latest set of installation instructions were given.
I suspect you still have a build problem related to the iconf_pgplot script.
see my previous message about not knowing where to place the new script that you gave us. Looking forward to hearing from you. Stefano -- ____________________________________________ Stefano Ciurli Professor of Chemistry Department of Agro-Environmental Science and Technology University of Bologna Viale Giuseppe Fanin, 40 I-40127 Bologna Italy Phone: +39-051-209-6204 Fax: +39-051-209-6203 "Fatti non foste a viver come bruti, ma per seguir virtute e canoscenza" Dante Alighieri - Inferno - Canto XXVI
participants (4)
-
Jeff Terry -
Matt Newville -
Paul Fons -
Stefano Ciurli