| delvar [message #14262] |
Thu, 04 February 1999 00:00  |
Dave Brennan
Messages: 28
Registered: December 1998
|
Junior Member |
|
|
Hi
I was wondering, is there another way of deleting variable to clear
memory other than using delvar. The probelm with delvar is that it
resets the main program, thus stopping the remainder from being
processed.
Thanks for your help
Dave Brennan
|
|
|
|
| Re: delvar [message #14317 is a reply to message #14262] |
Tue, 09 February 1999 00:00  |
thompson
Messages: 584
Registered: August 1991
|
Senior Member |
|
|
Alex Schuster <alex@rosa.mpin-koeln.mpg.de> writes:
> Dominic Zarro wrote:
>> The following may not guarantee that a variable
>> is permanently deleted from IDL memory, but it
>> is useful for undefining the variable:
>>
>> IDL> a = temporary(a)
> Um, you didn't actually try this, did you? ;-)
I think that what Dominic was trying to say here was that if you wrote a simple
routine along the lines
pro delete_variable, variable
dummy = temporary(variable)
end
you could then delete a variable by saying
delete_variable, a
I know that we've been doing that for a while now, using a similar procedure
that Dominic wrote. I don't know how this would work with pointer variables,
but it sounds like Dominic's on top of that, too.
Bill Thompson
|
|
|
|
| Re: delvar [message #14318 is a reply to message #14262] |
Tue, 09 February 1999 00:00 |
Martin Schultz
Messages: 515
Registered: August 1997
|
Senior Member |
|
|
Stein Vidar Hagfors Haugan wrote:
>
> [... snipped ]
> Which again reminds me - it would be *really* nice if RSI would
> start putting in "Not before version X.XX" information in the
> online documentation for (new) features. It's tiresome enough
> trying to write version-sensitive code, and I just don't need
> the extra work required to check exactly which of the current
> features go back to this or that version...
>
> Stein Vidar
>
How about decreasing the speed with which they release new versions
these days? This almost reminds me of Microsoft where you had just typed
in the first sentence in your new version of word and the next update
(or bug fix) came ... Personally, I would like to see some split between
interim versions and major version updates. There should always be a
latest official release which would remain stable for at least one year,
and there could be additional "sneak preview" releases for those who
want to try out new features at once (and get involved in beta or gamma
testing). Then there would at least be some official standard as to what
keywords are available, and everything else would be at own risk; even
with possible changes in the next big release (so they would have a
chance to correct inconsistencies without annoying too many people). And
for most of us, the
IF (!VERSION ...) block could be held at a reasonable size.
I do agree that listing the version range in the online help would
already help a lot! Even better, of course, if we could also find cross
links to the correct places for older or newer versions of what we
are/were used to.
Martin.
--
------------------------------------------------------------ -------
Dr. Martin Schultz
Department for Engineering&Applied Sciences, Harvard University
109 Pierce Hall, 29 Oxford St., Cambridge, MA-02138, USA
phone: (617)-496-8318
fax : (617)-495-4551
e-mail: mgs@io.harvard.edu
Internet-homepage: http://www-as.harvard.edu/people/staff/mgs/
------------------------------------------------------------ -------
|
|
|
|
|
|
| Re: delvar [message #14324 is a reply to message #14262] |
Tue, 09 February 1999 00:00 |
steinhh
Messages: 260
Registered: June 1994
|
Senior Member |
|
|
In article <36c05a04.0@208.206.112.5> "Dominic Zarro"
<zarro@tidalwave.net> writes:
> Actually, my last response was incorrect.
> A better way, using pointers is as follows:
>
> ;-- to delete a variable var
>
> ;-- allocate a pointer
>
> *p=ptr_new(/all)
Hi Dominic,
I guess you mean
p=ptr_new(/alloc)
(Seeing the "/all" in the ptr_new made me *really* confused,
I think the two extra letters are very clarifying :-)
[..]
>
> You can easily package the above into a simple
> procedure that deletes any variable var.
How about a one-line "syllogism":
ptr_free,ptr_new(temporary(var))
It's amazing what you have to do to delete a variable in IDL :-)
This discussion reminded me of a procedure I wrote some time in
1994, called like this:
ASSIGN_NOCOPY,destination,source
This was way before temporary() appeared on stage, and it even
(still) has checks to function (resorts to copying) with IDL
3.0, which didn't even have the NO_COPY keyword in the
WIDGET_CONTROL,GET_UVALUE=var mechanism....
Then came the handles, and then...
Which again reminds me - it would be *really* nice if RSI would
start putting in "Not before version X.XX" information in the
online documentation for (new) features. It's tiresome enough
trying to write version-sensitive code, and I just don't need
the extra work required to check exactly which of the current
features go back to this or that version...
Stein Vidar
|
|
|
|
| Re: delvar [message #14327 is a reply to message #14262] |
Mon, 08 February 1999 00:00 |
Dominic Zarro
Messages: 7
Registered: July 1998
|
Junior Member |
|
|
Actually, my last response was incorrect.
A better way, using pointers is as follows:
;-- to delete a variable var
;-- allocate a pointer
*p=ptr_new(/all)
;-- copy the variable into pointer
*p=temporary(var) ;-- temporary will remove var
;-- free the pointer
ptr_free,p
You can easily package the above into a simple
procedure that deletes any variable var.
Dominic
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
