| Re: Saving / Restoring Objects [message #35031] |
Fri, 09 May 2003 12:33  |
JD Smith
Messages: 850
Registered: December 1999
|
Senior Member |
|
|
On Fri, 09 May 2003 06:33:06 -0700, David Fanning wrote:
> mfeldt (mfeldt@mpia.de) writes:
>
>> trying myself and surfing the web I find a lot of remarks how useful
>> it is to save and restore idl objects. However, for me it doesn't
>> seem to work - am I doing something wrong here? Basically what I do is
>> this:
>>
>> IDL> a=obj_new('fancy_object')
>> % Compiled module: FANCY_OBJECT__DEFINE. ;; ;; this compiled my object
>> with lots of functions etc.. ;; IDL> a-> set,'Debug',1 ; internal
>> variable access... ;; ;; many more of those may go here ;; ;; then:
>> IDL> save,a,file='test.sav'
>>
>> IDL> exit
>>
>> ;; now a new session
>>
>> IDL> restore, file='test.sav'
>> IDL> help,a,/obj
>> ** Object class FANCY_OBJECT, 0 direct superclasses, 0 known methods
>>
>> ;; i.e. the object seems to be there, but all information on it and
>> all methods are lost... I also tried compiling the related code first,
>> but no use...
>>
>> Is there any way to make this work??
>
> I don't think the methods are "lost", they just haven't yet made
> themselves known to IDL. (Although I would have thought compiling the
> routines should have worked and would be essential--you use
> Resolve_Routine, right?--to restore the object properly) In any case, I
> would feel free to use your restored object as normal and see what
> happens.
>
> I guess whether the methods can be "found" will depend to some extent on
> how you are naming the methods and the files that contain them. But this
> SAVE/RESTORE method certainly works for objects.
>
>
If all your methods are in the fancy_object__define file, then you'll have
to explicitly compile this file. This is because the saved object
implicitly contains the class definition (but no methods). Hence, IDL
never feels the need to run the class definition procedure at the end of
fancy_object__define.pro, and your methods defined there remain hidden.
This problem has been discussed in great detail over the years, and David
even has a topic on his site outlining a workaround.
One alternative easy option is to make a separate file for each method,
e.g. fancy_object__init.pro. This gets fairly distracting fairly
quickly, so most people prefer the resolve_obj method.
JD
|
|
|
|
| Re: Saving / Restoring Objects [message #35040 is a reply to message #35031] |
Fri, 09 May 2003 06:33  |
David Fanning
Messages: 11724
Registered: August 2001
|
Senior Member |
|
|
mfeldt (mfeldt@mpia.de) writes:
> trying myself and surfing the web I find a lot of remarks how useful
> it is to save and restore idl objects. However, for me it doesn't
> seem to work - am I doing something wrong here? Basically what I do is
> this:
>
> IDL> a=obj_new('fancy_object')
> % Compiled module: FANCY_OBJECT__DEFINE.
> ;;
> ;; this compiled my object with lots of functions etc..
> ;;
> IDL> a-> set,'Debug',1 ; internal variable access...
> ;;
> ;; many more of those may go here
> ;;
> ;; then:
> IDL> save,a,file='test.sav'
>
> IDL> exit
>
> ;; now a new session
>
> IDL> restore, file='test.sav'
> IDL> help,a,/obj
> ** Object class FANCY_OBJECT, 0 direct superclasses, 0 known methods
>
> ;; i.e. the object seems to be there, but all information on it and
> all methods are lost... I also tried compiling the related code first,
> but no use...
>
> Is there any way to make this work??
I don't think the methods are "lost", they just haven't
yet made themselves known to IDL. (Although I would have
thought compiling the routines should have worked and would
be essential--you use Resolve_Routine, right?--to restore
the object properly) In any case, I would feel free to use
your restored object as normal and see what happens.
I guess whether the methods can be "found" will depend to
some extent on how you are naming the methods and the files
that contain them. But this SAVE/RESTORE method certainly
works for objects.
Cheers,
David
--
David W. Fanning, Ph.D.
Fanning Software Consulting, Inc.
Phone: 970-221-0438, E-mail: david@dfanning.com
Coyote's Guide to IDL Programming: http://www.dfanning.com/
Toll-Free IDL Book Orders: 1-888-461-0155
|
|
|
|
| Re: Saving / Restoring Objects [message #35108 is a reply to message #35031] |
Tue, 13 May 2003 03:20 |
mfeldt
Messages: 4
Registered: October 2002
|
Junior Member |
|
|
JD Smith <jdsmith@as.arizona.edu> wrote in message news:<pan.2003.05.09.19.33.29.826287.15802@as.arizona.edu>...
> On Fri, 09 May 2003 06:33:06 -0700, David Fanning wrote:
>
>> mfeldt (mfeldt@mpia.de) writes:
>>
>>> trying myself and surfing the web I find a lot of remarks how useful
>>> it is to save and restore idl objects. However, for me it doesn't
>>> seem to work - am I doing something wrong here? Basically what I do is
>>> this:
>>>
>>> IDL> a=obj_new('fancy_object')
>>> % Compiled module: FANCY_OBJECT__DEFINE. ;; ;; this compiled my object
>>> with lots of functions etc.. ;; IDL> a-> set,'Debug',1 ; internal
>>> variable access... ;; ;; many more of those may go here ;; ;; then:
>>> IDL> save,a,file='test.sav'
>>>
>>> IDL> exit
>>>
>>> ;; now a new session
>>>
>>> IDL> restore, file='test.sav'
>>> IDL> help,a,/obj
>>> ** Object class FANCY_OBJECT, 0 direct superclasses, 0 known methods
>>>
>>> ;; i.e. the object seems to be there, but all information on it and
>>> all methods are lost... I also tried compiling the related code first,
>>> but no use...
>>>
>>> Is there any way to make this work??
>>
>> I don't think the methods are "lost", they just haven't yet made
>> themselves known to IDL. (Although I would have thought compiling the
>> routines should have worked and would be essential--you use
>> Resolve_Routine, right?--to restore the object properly) In any case, I
>> would feel free to use your restored object as normal and see what
>> happens.
>>
>> I guess whether the methods can be "found" will depend to some extent on
>> how you are naming the methods and the files that contain them. But this
>> SAVE/RESTORE method certainly works for objects.
>>
>>
> If all your methods are in the fancy_object__define file, then you'll have
> to explicitly compile this file. This is because the saved object
> implicitly contains the class definition (but no methods). Hence, IDL
> never feels the need to run the class definition procedure at the end of
> fancy_object__define.pro, and your methods defined there remain hidden.
> This problem has been discussed in great detail over the years, and David
> even has a topic on his site outlining a workaround.
>
> One alternative easy option is to make a separate file for each method,
> e.g. fancy_object__init.pro. This gets fairly distracting fairly
> quickly, so most people prefer the resolve_obj method.
>
> JD
O.K., thanks folks .. that's it - quite easy once you know it. The
trick was actually to use ".compile" instead of ".r". After using IDL
quite extensively for almost 10 years now, I have hardly ever used
this ccommand before...
BTW: resolve_obj ... from which version on does this exist??
Thanks
Markus
--
Markus Feldt Voice: +49 6221 528 262
Max-Planck-Institut Fax: +49 6221 528 246
fA�r Astronomie mailto:mfeldt@mpia.de
KA�nigstuhl 17 http://www.mfeldt.de
D-69117 Heidelberg, Germany Si, !asi es la vida!
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
