SSW with GDL

[james-kc]

Hello, I am attempting to use the SSW (Solar SoftWare) library with GDL but I am running into issues.
Does anyone know if it is possible to use SSW with GDL?
Thanks!

Edit: Running Ubuntu 22.04.3 and GDL v1.0.1

[alaingdl]

Could you please report GDL version, and the OS you are using ? thanks

[james-kc]

I'm running Ubuntu 22.04.3 and GDL v1.0.1. Thanks

[alaingdl]

Very good news. Could you give more details on the issues ?
@GillesDuvert spend some time for SSW and has good results !

[james-kc]

That’s great! My issues are mostly trying to tell SSW to look for GDL rather than IDL. I have both GDL and SSW installed however I’m just having trouble getting them to talk to each other!

My understanding is that I should make a setup file that sets environment variables such as IDL_DIR, IDL_PATH, SSW_INSTR etc but I might be misunderstanding what each of these variables mean / what they should be set to for use with GDL.

Thanks!

[alaingdl]

please have a look to #1656

another solution, as a temporary work-around, is to use PATH_ADD procedure (it should be in the path)
to add recursively all the sub dirs of SSW but maybe it is not the normal way for SSW

$ gdl
GDL> path_add
% Compiled module: PATH_ADD.
USAGE : pro PATH_ADD, paths [,/current][,/recursive][,/prepend][,/append][,/help][,/verbose]
GDL> #path_add    ; # is a GDL extension to have details on internal or compiled func/pro
Compiled PROCEDURE : PATH_ADD,[1 Arg],TEST,VERBOSE,HELP,BEFORE,PREPEND,AFTER,APPEND,RECURSIVE,CURRENT
GDL> path_add, /pre,/rec, '/the/path/to/the/root/of/SSW'

HTH

A.

PS: for such a code, you should consider to add somewhere at the end of the GDL_PATH the path to the IDL proc. ...

[alaingdl]

I forgot how complex & big was SolarSoft ! 🥲

Does someone known some basic examples I can try ?
(I have a basic version in IDL & just adding paths in GDL ...)

[james-kc]

If this helps, this is the bash script I run to open SSW when I am using IDL in university:

#!/bin/tcsh
# simple script to run SSWIDL

# configure IDL
source /opt/harris/idl/bin/idl_setup

# configure the IDL_PATH
setenv IDL_PATH "+$HOME/idl::<IDL_DEFAULT>"

# add in the idllibs folders
setenv IDL_PATH "+/opt/idllibs/:$IDL_PATH"

# SSW environment variables
setenv SSW /opt/ssw
setenv SSW_INSTR "aia hmi eve sumer chianti eis eit cds trace mdi sot xrt iris hessi trace hxrs nso vso goes secchi spex xray yohkoh ontology stereo mjastereo"
source $SSW/gen/setup/setup.ssw /quiet

#actually run SSWIDL
sswidl

exit 0

I am essentially wishing to replicate this for use with GDL instead of IDL.

[GillesDuvert]

This is what I did

1. in ssw_idl, replace $IDL_DIR/bin/idl $command (line 316 in my version) by
   gdl $command
   and rename ssw_idl to 'ssw_gdl' .
2. add alias sswgdl $SSW/gen/setup/ssw_gdl after line containing alias sswidl $SSW/gen/setup/ssw_idl in ssw/gen/setup/setup.ssw

switch to tcsh (since ssw only has tcsh scripts), my .tcshrc contains

setenv SSW $HOME/ssw
setenv SSW_INSTR 'hessi xray spex'
source $SSW/gen/setup/setup.ssw

call ssw_gdl

and voila!

NB: possibly, redefine !PATH when inside ssw_gdl to contain /usr/local/share/gnudatalanguage/lib (as first elements of the path) if problems arise.

[GillesDuvert]

Not exactly the right moment as I'm busy with 'GDL exact plots', but I'll like hearing about the level of succes you have.
My tests were : hessi seems to work ok.
ssw is a real challenge, so do not be suprised if something goes wrong.

[GillesDuvert]

Ah, @james-kc : 1.0.1 may be too old.At least 1.0.3 if any, but it would be better with current version (GitHub development).

[james-kc]

Thanks for your help @GillesDuvert! I think I am close to getting this to work however I am getting an error message about IDL_DIR.

This is the contents of a gdl startup script which I have placed in /usr/local/bin to allow me to use gdl $command in $SSW/gen/setup/ssw_gdl:

#!/bin/bash
#
# must be run as a BASH Shell
#
# AC: improved shell to avoid "big" when sourcing (or ".") this file ...
# echo "(ba)sh or source ? "$BASH_SOURCE
#
# AC 01/10/2010: we can keep the current $GDL_PATH using -g [--gdl_path]
#

if [ $BASH_SOURCE != $0 ] ; then
   echo "Please run this script throw one in the 2 next commands:"
   echo "1/    ./quick_run_GDL.sh"
   echo "2/    bash ./quick_run_GDL.sh"
   return
fi
#
# AC 2011/08/19 to avoid conflict with other IDL/GDL environmments
# which have been setup before
#
REF_IDL_PATH=
REF_GDL_PATH=
REF_IDL_STARTUP=
REF_GDL_STARTUP=
#
IDL_PATH=""
GDL_PATH=""
IDL_STARTUP=""
GDL_STARTUP=""
#
if [ -z $flag_help ] ;     then flag_help=0 ; fi
if [ -z $flag_quiet ] ;    then flag_quiet=0 ; fi
if [ -z $flag_valgrind ] ; then flag_valgrind=0 ; fi
if [ -z $flag_gdl_path ] ; then flag_gdl_path=0 ; fi
# 
while getopts ":ghqv-:" opt ; do
    case $opt in 
	- ) case $OPTARG in
	    quiet* )    echo "Quiet keyword swith ON"
	                flag_quiet=1 ;;
	    gdl_path)   echo "Keeping GDL_PATH swith ON"
	                flag_gdl_path=1 ;;
	    help* )     echo "Help keyword swith ON"
                        flag_help=1 ;;
	    valgrind* ) echo "Valgrind keyword swith ON"
                        flag_valgrind=1 ;;
	    ? )         echo "option illegal -$OPTARG
	                exit ;;
	    * )         echo "option illegal -$OPTARG
	                exit
	esac ;;
    q )  echo "Quiet keyword swith ON"
         flag_quiet=1 ;;
    g )  echo "Keeping GDL_PATH swith ON"
         flag_gdl_path=1 ;; 
    h )  echo "Help keyword swith ON"
         flag_help=1 ;;
    v* ) echo "Valgrind keyword swith ON"
         flag_valgrind=1 ;;
    ? )  echo "option illegal -$OPTARG"
         exit ;;
    esac
done
#
if [ "$flag_help" -eq 1 ] ; then
    echo "Usage: gdl -g [--gdl_path] -h [--help] -q [--quiet] -v [--valgrind]"
    exit
fi
#
if [ "$flag_quiet" -eq 0 ] ; then
    echo "Running current GDL with CMAKE_SOURCE_DIR/src/pro and CMAKE_SOURCE_DIR/testsuite in GDL_PATH"
fi
#
if [ "$flag_gdl_path" -eq 0 ] ; then
    export GDL_PATH='+/home/james/GDL/gdl/src/pro/:+/home/james/GDL/gdl/testsuite/'
else
    export GDL_PATH='+/home/james/GDL/gdl/src/pro/:+/home/james/GDL/gdl/testsuite/:'
fi
#
if [ "$flag_quiet" -eq 0 ] ; then
    echo $GDL_PATH
fi
#
if [ "$flag_valgrind" -eq 1 ] ; then
    valgrind /home/james/GDL/gdl/build/src/gdl
else
    /home/james/GDL/gdl/build/src/gdl
fi
#
# restoring initial envionment
#
export IDL_PATH=
export GDL_PATH=
export IDL_STARTUP=
export GDL_STARTUP=

I have created this file to allow me to run the tcsh commands easily:

#!/bin/tcsh

setenv SSW /usr/local/ssw
setenv SSW_INSTR 'aia hmi eve sumer chianti eis eit cds trace mdi sot xrt iris hessi trace hxrs nso vso goes secchi spex xray yohkoh ontology stereo mjastereo'
source $SSW/gen/setup/setup.ssw

sswgdl

When this file is called, I get the following response:

SSW  setup will include: <gen aia hmi eve sumer chianti eis eit cds trace mdi sot xrt iris hessi trace hxrs nso vso goes secchi spex xray yohkoh ontology stereo mjastereo>

Type <sswidl> to start SSW IDL
Cannot find idl directory - (IDL_DIR)

I imagine it may be that I have not correctly configured the gdl startup script correctly but I am not sure where to go from here.
Thanks!

[james-kc]

P.S. I updated to GDL Version 1.0.3 Git before trying the above!

[james-kc]

P.P.S. Another possibility is that I have not installed GDL correctly!

[alaingdl]

Did you make a "make install" or "build_gdl.sh install" after the compilation ? This should solve the message :

*** PLPLOT ERROR, ABORTING OPERATION ***
plInitDispatchTable: Could not open drivers directory, aborting operation

On my side, when I compile without the script, I do :

(at the root of the cloned GDL ...)
git clone https://github.com/gnudatalanguage/gdl
cd gdl
git submodule update --init
mkdir mybuild
cd mybuild
cmake .. -DQHULL=OFF -DCMAKE_INSTALL_PREFIX=$PWD/here
make install
cd ..
./mybuild/here/bin/gdl

and no plplot message 😃

[GillesDuvert]

@james-kc the errors messages at start of gdl mean that our 'modified' drivers (for the plplot library) are not found. As @alaingdl mentioned, this should not happen if gdl is properly installed (calling build_gdl.sh install after build_gdl.sh build).

[james-kc]

That’s what I’ve missed!

To install GDL, I used "way 1" detailed here:

https://github.com/gnudatalanguage/gdl/wiki/GDL-on-Linux

I missed the final build_gdl.sh install step so hopefully that solves my issue. I’ll have a go at this later today. Thanks!

[GillesDuvert]

concerning IDL_DIR:

• the ssw_idl script included in the ssw distribution does, I quote: " Verify a working IDL_DIR is found ". As I have IDL installed on my machine (free demo version), it does not complain for me. You on the contrary have no IDL, so you need to insure that this part of ssw_idl (ssw_gdl in your case) is not active and probably also that IDL_DIR is defined to something, just in case.
• I never use the gdl startup script of AC above, as a properly installed gdl behaves ok (just call 'gdl'). The above script is useful for maintaining several GDL / IDL environments, that is not your case. Besides, ssw_idl and the ssw people took great care that everyting in the IDL environment is OK before allowing ssw to run. This makes our task quite difficult to start with, no need to add another dependency through the use of this above script.

[alaingdl]

using the modifications mentioned by Gilles in #1659 (reply in thread) and a personnel csh script I am able to start SSW ... (where are the demo files ?!)

#!/bin/csh
setenv GDL_LIB "+/home/alaingdl/GDL/gdl-1.0.3git231020/mybuild/ici/share/gnudatalanguage/lib/"
setenv MYHOME /home/alaingdl
setenv SSW $MYHOME/ssw/
setenv PATH "$PATH":$SSW
setenv SSW_INSTR "gen eit hessi secchi stereo vobs ontology swap"
source $SSW/gen/setup/setup.ssw /quiet
setenv IDL_DIR /home/alaingdl/IDLetGDL/IDL88/idl/
setenv IDL_PATH $IDL_DIR/lib:+$MYHOME/IDLetGDL/Librairies/DebianAstroGIT/
setenv IDL_PATH "$GDL_LIB":"$IDL_PATH"
setenv WORKING_PATH ./
sswgdl

and yes, I made a link to the bin dir. where is "gdl" in the ssw/ path ...

Not super clean but working so far ...

[GillesDuvert]

yes if there were test procedures on demo files for each instrument, it would be great to run them!

[james-kc]

I have finally been able to get SSW working with GDL! Thank you everyone for your help!

[james-kc]

One final question now; does GDL have the following functions available?
STRSPLIT
STRMATCH
LMGR
Not sure if this is another issue of me not installing GDL correctly or whether these are actually part of a library (although as far as I know, these functions are build into IDL)
Thanks!

[alaingdl]

Those routines are available in GDL, and this is why my lines :

setenv GDL_LIB "+/home/alaingdl/GDL/gdl-1.0.3git231020/mybuild/ici/share/gnudatalanguage/lib/"
setenv IDL_PATH "$GDL_LIB":"$IDL_PATH"

are so important !

"bon courage"

A.

[GillesDuvert]

Adding @alaingdl remark, this could be a more complete answer :

• in ssw_idl:

1. add setenv GDL_LIB "+/usr/local/share/gnudatalanguage/lib/" after # set the darn thing. (note the + sign)
2. replace setenv IDL_PATH $TEMP2 by setenv IDL_PATH "$GDL_LIB":"$TEMP2" (note the quotes)
3. replace$IDL_DIR/bin/idl $command by gdl $command
4. save the file as 'ssw_gdl' .

• add alias sswgdl $SSW/gen/setup/ssw_gdl after line containing alias sswidl $SSW/gen/setup/ssw_idl in ssw/gen/setup/setup.ssw

switch to tcsh (since ssw only has tcsh scripts)
call ssw_gdl

in the above, the location of the gdl procedures (library) is the default one, /usr/local/share/gnudatalanguage/lib/ . You (or your distribution maintainer) may have installed gdl elsewhere.

Plaese note that you must have curl installed for ssw getting remote data and updates.

[james-kc]

I have tried both what yourself and Alain have suggested but I am not getting any success. These are the lines I have in ssw_gdl

# set the darn thing.
setenv GDL_LIB "+/home/james/GDL/gdl/mybuild/here/share/gnudatalanguage/lib/"
setenv IDL_PATH "$GDL_LIB":"$TEMP2"
alias idltool $IDL_DIR/bin/idltool

# call IDL with any arguments
onintr -                                # allow ctrl c in idl
# gdl $command          # no alias (see comment) slf 30-Jul
 ~/GDL/gdl/mybuild/here/bin/gdl $command

When I run ssw using the following script:

#!/bin/tcsh

setenv MYHOME /home/james
setenv SSW /usr/local/ssw
setenv PATH "$PATH":$SSW
setenv SSW_INSTR 'aia hmi eve sumer chianti eis eit cds trace mdi sot xrt iris hessi trace hxrs nso vso goes secchi spex xray yohkoh ontology stereo mjastereo'
source $SSW/gen/setup/setup.ssw
setenv IDL_DIR /

sswgdl

I get this output:

SSW  setup will include: <gen aia hmi eve sumer chianti eis eit cds trace mdi sot xrt iris hessi trace hxrs nso vso goes secchi spex xray yohkoh ontology stereo mjastereo>

Type <sswidl> to start SSW IDL
--------------------------------------------------------
Running SSW, Last Updated: Wed Nov 2 22:22:51 GMT 2005
 
PROBLEMS? - e-mail TO: [email protected]
--------------------------------------------------------
  GDL - GNU Data Language, Version 1.0.3 Git
- For basic information type HELP,/INFO
- Default library routine search path used (GDL_PATH/IDL_PATH env. var. not set): +/home/james/GDL/gdl/mybuild/here/share/gnudatalanguage/lib/:+/usr/local/ssw/site/idl:/usr/local/ssw/gen/idl_test/conflicts_resolve:+/usr/local/ssw/gen/idl_test/genutil:/usr/local/ssw/gen/idl/atest:+/usr/local/ssw/gen/idl:/usr/local/ssw/sdo/aia/idl/atest:+/usr/local/ssw/sdo/aia/idl:/usr/local/ssw/sdo/hmi/idl/atest:+/usr/local/ssw/sdo/hmi/idl:/usr/local/ssw/sdo/eve/idl/atest:+/usr/local/ssw/sdo/eve/idl:/usr/local/ssw/hinode/eis/idl/atest:+/usr/local/ssw/hinode/eis/idl:/usr/local/ssw/hinode/sot/idl/atest:+/usr/local/ssw/hinode/sot/idl:/usr/local/ssw/hinode/xrt/idl/atest:+/usr/local/ssw/hinode/xrt/idl:/usr/local/ssw/iris/idl/atest:+/usr/local/ssw/iris/idl:/usr/local/ssw/hessi/idl/atest:+/usr/local/ssw/hessi/idl:/usr/local/ssw/packages/spex/idl/atest:+/usr/local/ssw/packages/spex/idl:/usr/local/ssw/packages/xray/idl/atest:+/usr/local/ssw/packages/xray/idl:/usr/local/ssw/vobs/ontology/idl/atest:+/usr/local/ssw/vobs/ontology/idl:+//lib:+//userlib:+/usr/local/ssw/gen/idl_libs/astron
- Using WxWidgets as graphics library (windows and widgets).
- Using local drivers in /home/james/GDL/gdl/mybuild/here/bin/../share/gnudatalanguage/drivers
- Please report bugs, feature or help requests and patches at: https://github.com/gnudatalanguage/gdl

Executing SSW IDL_STARTUP for: GEN
Executing SSW IDL_STARTUP for: AIA
% Compiled module: REM_ELEM.
% Compiled module: REM_ELEM2.
% Compiled module: GET_UNIQ.
% Compiled module: REMOVE.
% Compiled module: SSW_UNIQ.
% Compiled module: DEFAULT.
% Compiled module: DATATYPE.
% VALUE_LOCATE: STRING expression not allowed in this context:RM
% SSW_PATH: Procedure not found: PRSTR
% Procedure not found: SET_LOGENV
% SSW_PATH: Procedure not found: PRSTR
Executing SSW IDL_STARTUP for: HMI
% SSW_PATH: Procedure not found: PRSTR
% Procedure not found: SET_LOGENV
% SSW_PATH: Procedure not found: PRSTR
Executing SSW IDL_STARTUP for: EVE
% SSW_PATH: Procedure not found: PRSTR
% Procedure not found: SET_LOGENV
% SSW_PATH: Procedure not found: PRSTR
Executing SSW IDL_STARTUP for: EIS
% Procedure not found: LOAD_EXTERNAL
% Ambiguous: Variable is undefined: FIX_PATH or: Function not found: FIX_PATH
% Ambiguous: Variable is undefined: FIX_PATH or: Function not found: FIX_PATH
% Ambiguous: Variable is undefined: FIX_PATH or: Function not found: FIX_PATH
% Procedure not found: EIS_DBASE
Executing SSW IDL_STARTUP for: SOT
% SSW_PATH: Procedure not found: PRSTR
% SSW_PATH: Procedure not found: PRSTR
Executing SSW IDL_STARTUP for: XRT
% SSW_PATH: Procedure not found: PRSTR
% Procedure not found: SET_LOGENV
Executing SSW IDL_STARTUP for: IRIS
% SSW_PATH: Procedure not found: PRSTR
% Procedure not found: SET_LOGENV
% Procedure not found: SET_LOGENV
% Procedure not found: RECOMPILE
% Procedure not found: IRIS_MSU_CASE
% SSW_PATH: Procedure not found: PRSTR
Executing SSW IDL_STARTUP for: ONTOLOGY
% Procedure not found: RECOMPILE
Executing SSW IDL_STARTUP for: SITE
% SSW_PATH: Procedure not found: UCON_PATH
% Ambiguous: Variable is undefined: IS_MEMBER or: Function not found: IS_MEMBER

GDL> goes_flare_list('2017-09-10', '2017-09-11')

% Compiled module: GOES_FLARE_LIST.
% GOES_FLARE_LIST: Function not found: OGOES
% Execution halted at: GOES_FLARE_LIST     26 /home/james/Documents/sff_experimentation/goes_flare_list.pro
%                      $MAIN$

GDL> .r 

% Error opening file. File: strsplit. 

However, before I added GDL_LIB and ammended the IDL_PATH in ssw_gdl, the OGOES function was working while the STRSPLIT function also did not work.
I am wondering have I still not installed GDL correctly or have I missed a part of the SSW installation where I setup pointers to the SSW libraries (although, it does not seem to find the functions even when I do

setenv GDL_LIB "+/home/james/GDL/gdl/mybuild/here/share/gnudatalanguage/lib/"
setenv IDL_PATH "$GDL_LIB":"$TEMP2"

And just to double check, this is the contents of /home/james/GDL/gdl/mybuild/here/share/gnudatalanguage/lib/

[GillesDuvert]

Very probably you installed gdl correctly.
You can test it outside the ssw environment, in bash in a new session, just call
gdl
And type
plot, dist(10)
If a plot appears then your graphic drivers are well configured and the default !PATH is correct.

[GillesDuvert]

The above problems just mean that the SSW libraries are not found in the !PATH when gdl starts. Meaning that the IDL_PATH does not contain ssw libraries anymore. So there is a problem in the script edits described above.

[GillesDuvert]

And, again, please just install gdl ( build_gdl.sh install), refer to gdl as 'gdl' ( without path) and do not use any gdl startup script. Just 'ssw_gdl' at the tcsh prompt. Every other way is too confusing.

[james-kc]

Ahh, I haven't been able to get gdl to run just by typing gdl, I have only been able to run it using ~/GDL/gdl/mybuild/here/bin/gdl.
After installing GDL using the build_gdl.sh method (including the build_gdl.sh install command), I was still not able to call it using gdl so I tried the method described in this reply: #1659 (reply in thread)

Also, in tcsh, I am not able to call ssw_gdl just by typing ssw_gdl.

[jeget]

Thanks for the good news! I started testing SSW under GDL 1.0.3, but first I'd like to ask a few questions about installation.
How do I uninstall the older version? For now, I renamed the existing gdl directory and git-cloned a new one.
I installed successfully, but several tests failed:

The following tests FAILED:
	 81 - test_dynamic_drivers.pro (Failed)
	105 - test_hdf5.pro (Failed)
	147 - test_pmulti.pro (Failed)
	149 - test_postscript.pro (Failed)
Errors while running CTest
make: *** [Makefile:71: test] Error 8

Is it ok?
And finally, as far as I understood, I should start GDL by simply typing gdl. However, right after installation I cannot do this, I only have to type the full path. Is it wrong?

[alaingdl]

I suspect 81 is related to a path issue. Did you made "make install" or "build_gdl.sh install" before running the test ?

105 is just one HDF5 related code (outside GDL) is not in the path (not installed !) A detail

147 & 149 are related to recent changes in the code @GillesDuvert

you have to locate the binary ($ find . -name gdl) where you made the compilation, and start with the path related to it
(eg ./install/bin/gdl on my laptop)

Would be great (and cautious) to removed packaged version ! A risk it will removed few dependences. Then play again "build_gdl.sh prep" to re-installed them !

HTH
A.

[alaingdl]

You should have something like :

/bin/bash: line 1: h5diff: command not found

when you run test_hdf5 procedure in GDL (assuming you have testsuite/ path in the GDL path !)
(the same in IDL !)

As you can see in the test_hdf5.pro, we test using external h5diff. In the Debian & Ubuntu universe :

$ LANG=C h5diff
Command 'h5diff' not found, but can be installed with:
sudo apt install hdf5-tools

[alaingdl]

Concerning being in the PATH, it can be opening a long discussion. Today a lot of colleagues do have Conda related stuffs in their PATH and they have a lot of troubles with normal codes (not related to Conda (or Conda dependences) because those path are prepended by default to paths related to the system !

[jeget]

I reinstalled GDL, now hdf5 error is gone. I can start GDL by specifying the path so it works. I tried to test SSW once again. I encountered two problems: one with a graphical interface and one within command line interface.

1. When I type hessi, it starts a GUI, but in the terminal I see a message

GDL> 02:23:35: Debug: window wxMenuBar(0x55ea92d1c420, ) lost focus even though it didn't have it

I select File -> Select Observation Time Interval and write the start and end times in the upper line boxes, e.g. 17-Feb-2013 15:35:00.000 and 17-Feb-2013 16:00:00.000. Then I click Plot Observing Summary Data and a plot shows up. Now, if I click the left or right mouse button on the plot, something crashes. The terminal writes

Unknown user value in plotman =

whenever I move the mouse, and the graphical windows are lost, i.e. I can press buttons, but the windows do not react, and it is impossible to close them, even by closing the terminal. In IDL, when I press the left button over the plot, nothing happens. When I press the right button, the point's coordinates show up on the plot and are printed in the terminal.

2. I have a small script (note that it downloads some 450 MB of data)

search_network,/enable
obj=hsi_spectrum()
obj->set, decimation_correct=1
obj->set, obs_time_interval=['6-Jan-2004 06:00:00','6-Jan-2004 06:35:00']
obj->set,pileup_correct=0
obj->set,seg_index_mask=[1,0,1,1,0,1,0,1,1,0,0,0,0,0,0,0,0,0]
obj->set, sp_data_unit='Flux'
obj->set, sp_energy_binning=22
obj->set, sp_semi_calibrated=0
obj->set, sp_time_interval=4
obj->Set, sum_flag=1
obj->set, use_flare_xyoffset=1
obj->filewrite, /fits, /buildsrm, srmfile='srm.fits', specfile='spec.fits', all_simplify=0, /create

It works OK until the last line where I get the following (after some lines of output)

% Compiled module: RESTGEN.
% MAKE_STR: Parser syntax error: unexpected token: FLTARR
% TAG_NAMES: Error: Failed to obtain structure. Input type: STRING
% Error occurred at: BUILD_STR           63 /home/bulat/ssw/gen/idl/genutil/build_str.pro
%                    RD_GENX             66 /home/bulat/ssw/gen/idl/io/rd_genx.pro
%                    RESTGEN             63 /home/bulat/ssw/gen/idl/io/restgen.pro
%                    XR_RD_ABUNDANCE    150 /home/bulat/ssw/packages/xray/idl/xr_rd_abundance.pro
%                    SET_XR_ABUNDANCE    92 /home/bulat/ssw/packages/xray/idl/set_xr_abundance.pro
%                    CHIANTI_KEV_COMMON_LOAD   128 /home/bulat/ssw/packages/xray/idl/chianti_kev_common_load.pro
%                    CHIANTI_KEV_TEMP_RANGE    19 /home/bulat/ssw/packages/xray/idl/chianti_kev_temp_range.pro
%                    F_VTH              131 /home/bulat/ssw/packages/xray/idl/f_vth.pro
%                    F_VTH_BPOW          69 /home/bulat/ssw/packages/xray/idl/f_vth_bpow.pro
%                    HSI_DECIM_CORRECTION   126 /home/bulat/ssw/hessi/idl/spectra/hsi_decim_correction.pro
%                    HSI_SPECTROGRAM::DECIM_TABLE  2470 /home/bulat/ssw/hessi/idl/spectra/hsi_spectrogram__define.pro
%                    HSI_SPECTROGRAM::DECIM_CORRECT  2271 /home/bulat/ssw/hessi/idl/spectra/hsi_spectrogram__define.pro
%                    HSI_SPECTRUM::PROCESS_HOOK_POST  1001 /home/bulat/ssw/hessi/idl/spectra/hsi_spectrum__define.pro
%                    HSI_SPECTROGRAM::PROCESS  1125 /home/bulat/ssw/hessi/idl/spectra/hsi_spectrogram__define.pro
%                    FRAMEWORK::GETDATA   839 /home/bulat/ssw/gen/idl/objects/framework/framework__define.pro
%                    HSI_SPECTROGRAM::GETDATA   511 /home/bulat/ssw/hessi/idl/spectra/hsi_spectrogram__define.pro
%                    HSI_SPECTRUM::GETDATA   639 /home/bulat/ssw/hessi/idl/spectra/hsi_spectrum__define.pro
%                    HSI_SPECTRUM::SPECTRA4SPEX   109 /home/bulat/ssw/hessi/idl/spectra/hsi_spectrum__spectra4spex.pro
%                    HSI_SPECTRUM::FILEWRITE   283 /home/bulat/ssw/hessi/idl/spectra/hsi_spectrum__filewrite.pro
%                    $MAIN$          
% Execution halted at: XR_RD_ABUNDANCE    150 /home/bulat/ssw/packages/xray/idl/xr_rd_abundance.pro

[GillesDuvert]

@james-kc all this is very interesting, even if very frustrating for you.
I can (last version of GDL, soon published as 1.0.4, but this is probably irrelevant):

• start ssw using the ssw_gdl discussed above, no need to redefine !PATH (but my environment is not yours).
• type 'hessi' and get the gui (without the message you report ---> this is probably a trick of wxWidgets being compiled in debug mode (?) )
• cannot get the plot data, but that's because I've not downloaded/updated hessi data, it says;
• "% HSI_QLOOK::GETDATA: file: hsi_obssumm_20130217_055.fits NOT found"
• can however 'Plot GOES' although it complains about " RD_GOES_SDAC: No SDAC/GOES data available for specified times." and

 % Compiled module: STRJUSTIFY.
 ------------------------------------
| Problem parsing times in filenames |
 ------------------------------------

Which looks like problematic to me: normal, not normal?

Then I started your script, but again I have no data:

% Compiled module: REPRODUCE.
% HSI_PACKET_FILE::PROCESS: No files are available for this observing time interval. Aborting.
% Execution halted at: HSI_PACKET_FILE::PROCESS    42 /home/gildas/ssw/hessi/idl/atest/hsi_packet_file__define.pro

there may be a magic command to get the relevant data imported, or do we have a problem with the strings defining time intervals?

ssw is full of very tricky code, the one at line 63 in build_str.pro you report is a particularly convolved example.

[jeget]

@GillesDuvert yes, I use ssw_gdl in a pretty much the same way. I only replaced $IDL_DIR/bin/idl $command with $IDL_DIR/bin/gdl $command where IDL_DIR is /home/bulat/gdl/install, I did not change !IDL_PATH.
I forgot to say that before starting hessi you should type search_network, /enable. This allows the software to look for RHESSI data online. Plot GOES plots data from a different instrument, apparently it does not need that search_network.
However, in the second example I didn't forget it. So didn't you miss search_network, /enable in the beginning of the script?

[GillesDuvert]

indeed with search_network, /enable the plot came easily:

[jeget]

Great! What happens if you left- or right-click at some point of the plot?

[GillesDuvert]

sorry, bedtime now 😄

[GillesDuvert]

@jeget , I confirm 2 bugs:

1. when right-clicking on the plot, plots the date/position once, then runs in an infinite loop saying "Unknown user value in plotman =" ---> obviously a bug in the gdlwidget's code.
   idem when left-clicking (should open a zoom rectangle), same error.

2. above procedure: I confirm the error

% TAG_NAMES: Error: Failed to obtain structure. Input type: STRING
% Error occurred at: BUILD_STR           63 /home/gildas/ssw/gen/idl/genutil/build_str.pro

in GDL

These bugs have to be patched in the GDL code if you want to go further, I'm afraid.
As always with ssw, which consists in oodles of small procedures calling themselves via opaque 'objects', it may not be so easy to backtrace the origin of the problem, where the '=' comes from in the first case, and why the recursive function build_str is in error in the second case.
I suggest to close this discussion and pursue on the 2 issues I have opened #1685 and #1686

[alaingdl]

bad news on my side, with current git version, after adding hessi in ssw
ssw_upgrade,/hessi,/wget,/spawn, if I do hessi or obj=hsi_spectrum() I do have a :
Segmentation fault (core dumped)

Yes, independently, the widgets are OK (tested with test_widgets.pro)

[GillesDuvert]

At the moment, 2 of 3 bugs found (yes, there was anotehr) are closed by new PRs. Stay tuned.

[GillesDuvert]

Unfortunately, the procedure finally hangs on the absence of SPRSIN in GDL:

% HSI_MATMULT: Function not found: SPRSIN
% Execution halted at: HSI_MATMULT         61 /home/gildas/ssw/hessi/idl/spectra/hsi_matmult.pro
%                      HESSI_BUILD_SRM    510 /home/gildas/ssw/hessi/idl/spectra/hessi_build_srm.pro
%                      HSI_DECIM_CORRECTION   130 /home/gildas/ssw/hessi/idl/spectra/hsi_decim_correction.pro
%                      HSI_SPECTROGRAM::DECIM_TABLE  2470 /home/gildas/ssw/hessi/idl/spectra/hsi_spectrogram__define.pro
%                      HSI_SPECTROGRAM::DECIM_CORRECT  2271 /home/gildas/ssw/hessi/idl/spectra/hsi_spectrogram__define.pro
%                      HSI_SPECTRUM::PROCESS_HOOK_POST  1001 /home/gildas/ssw/hessi/idl/spectra/hsi_spectrum__define.pro
%                      HSI_SPECTROGRAM::PROCESS  1125 /home/gildas/ssw/hessi/idl/spectra/hsi_spectrogram__define.pro
%                      FRAMEWORK::GETDATA   839 /home/gildas/ssw/gen/idl/objects/framework/framework__define.pro
%                      HSI_SPECTROGRAM::GETDATA   511 /home/gildas/ssw/hessi/idl/spectra/hsi_spectrogram__define.pro
%                      HSI_SPECTRUM::GETDATA   639 /home/gildas/ssw/hessi/idl/spectra/hsi_spectrum__define.pro
%                      HSI_SPECTRUM::SPECTRA4SPEX   109 /home/gildas/ssw/hessi/idl/spectra/hsi_spectrum__spectra4spex.pro
%                      HSI_SPECTRUM::FILEWRITE   283 /home/gildas/ssw/hessi/idl/spectra/hsi_spectrum__filewrite.pro
%                      $MAIN$  

I have opened an issue, but I'm afraid this is not going to be solved soon. there are 4 needed routines to use sparse matrixes, namely SPRSAB, SPRSAX, SPRSIN and SPRSTP. All copyright-protected because published by Knuth. I note that the gnu scientific library has equivalent routines, and Eigen:: also . Adding an interface for these routines would be easy for the knowledgeable people. I open a request for them, but cannot do more.

[GillesDuvert]

In fact it's very easy to get all SPRxxx functions with Eigen::SparseMatrix in the background. A PR will be submitted soon.

[GillesDuvert]

HSI_MATMULT passed with the introduction of SPRSIN etc in GDL (#1698).
Now I get an error at:

% Compiled module: HSI_MATMULT.
% Compiled module: HSI_RM_BLANKET.
% Compiled module: HSI_RM_DETMOD.
% Compiled module: HSI_RM_INTERP_ROWS.
% Compiled module: HSI_RM_INTERP_READFILE.
% Compiled module: HSI_RM_INTERP_FIT.
% Compiled module: HSI_CORRECT_PEAKS.
% Compiled module: HSI_RM_FIT_EBOUNDARIES.
% Compiled module: HSI_RM_INTERP_CALC.
% Compiled module: HSI_REBINNER.
% Compiled module: HSI_RM_GRID2SCAT.
% Compiled module: HSI_MATADD.
% Compiled module: HSI_RM_SPACECRAFT.
% Compiled module: HSI_RM_EARTH.
% Compiled module: HSI_RM_DETRESOL.
% Compiled module: HSI_CALC_RESOLUTIONS.
% Compiled module: HSI_RM_ELECRESOL.
% Compiled module: HSI_RM_LLD.
% Compiled module: HSI_ATTEN_COEFF.
% Compiled module: READ_MAT_XCOM.
% Compiled module: COEFDATA.
% Compiled module: XCOM__DEFINE.
% Compiled module: XCOM_CONTROL.
% Compiled module: XCOM_CONTROL__DEFINE.
% Compiled module: XCOM_INFO__DEFINE.
% Compiled module: ELEMENT2Z.
% Compiled module: HSI_RD_SHUTTER_PARAMETERS.
% HSI_DECIM_CORRECTION: Subscript out of range [i].  Aborting decimation correction.
% Execution halted at: HSI_DECIM_CORRECTION    84 /home/gildas/ssw/hessi/idl/spectra/hsi_decim_correction.pro
%                      HSI_SPECTROGRAM::DECIM_TABLE  2470 /home/gildas/ssw/hessi/idl/spectra/hsi_spectrogram__define.pro
%                      HSI_SPECTROGRAM::DECIM_CORRECT  2271 /home/gildas/ssw/hessi/idl/spectra/hsi_spectrogram__define.pro
%                      HSI_SPECTRUM::PROCESS_HOOK_POST  1001 /home/gildas/ssw/hessi/idl/spectra/hsi_spectrum__define.pro
%                      HSI_SPECTROGRAM::PROCESS  1125 /home/gildas/ssw/hessi/idl/spectra/hsi_spectrogram__define.pro
%                      FRAMEWORK::GETDATA   839 /home/gildas/ssw/gen/idl/objects/framework/framework__define.pro
%                      HSI_SPECTROGRAM::GETDATA   511 /home/gildas/ssw/hessi/idl/spectra/hsi_spectrogram__define.pro
%                      HSI_SPECTRUM::GETDATA   639 /home/gildas/ssw/hessi/idl/spectra/hsi_spectrum__define.pro
%                      HSI_SPECTRUM::SPECTRA4SPEX   109 /home/gildas/ssw/hessi/idl/spectra/hsi_spectrum__spectra4spex.pro
%                      HSI_SPECTRUM::FILEWRITE   283 /home/gildas/ssw/hessi/idl/spectra/hsi_spectrum__filewrite.pro
%                      $MAIN$  

[jeget]

Thank you @GillesDuvert . I think I could try some other tests with SSW. In case of issues is it better to write here or open new ones?

[GillesDuvert]

@jeget that would be great! Please pursue here, the title of the discussion says all.

[GillesDuvert]

HESSI panels clicking will work now with #1694