| Re: RSI and the 24. Tagname... [message #8625] |
Mon, 24 March 1997 00:00 |
handy
Messages: 4
Registered: November 1996
|
Junior Member |
|
|
Joseph B. Gurman <gurman@ari.net> writes:
[...]
> Under 5.0b5, about which I guess it's Ok to speak in public because
> it;s a _public_ pre-release, I get:
> % Program code area full.
> Execute failed with 39Tagnames!!!
>
> under OpenVMS 7.1, and somewhat more spectacularly with 32 tagnames under
> DU4.0B.
>
> Interestingly, on a PowerMacintosh running MacOS 7.6, I get:
>
> Execute failed with 51Tagnames!!!
> So if nothing relating to structures is "improved" before the final
> release of 5.0, getting a PowerMac may be your only choice.
...Or maybe a Linux box:
(Running the posted 'fail.pro' code, IDL 5.0B5 on a FreeBSD-2.2 Box
running the linuxulator)
% Program code area full.
Execute failed with 249Tagnames!!!
IDL>
Happy Trails...
Brian
|
|
|
|
| Re: RSI and the 24. Tagname... [message #8629 is a reply to message #8625] |
Mon, 24 March 1997 00:00  |
steinhh
Messages: 260
Registered: June 1994
|
Senior Member |
|
|
In article <gurman-2303972309500001@goodgulf.nascom.nasa.gov>, gurman@ari.net (Joseph B. Gurman) writes:
|>
[.....]
|> Under 5.0b5, about which I guess it's Ok to speak in public because
|> it;s a _public_ pre-release, I get:
|>
|> .
|> .
|> .
|> Execute was OK with 37 Tagnames...
|> Execute was OK with 38 Tagnames...
|>
|> & END
|> ^
|> % Program code area full.
|> Execute failed with 39Tagnames!!!
|>
|> under OpenVMS 7.1, and somewhat more spectacularly with 32 tagnames under
|> DU4.0B.
|>
|> Interestingly, on a PowerMacintosh running MacOS 7.6, I get:
|>
|> Execute failed with 51Tagnames!!!
|>
|> but I don't think I can tell you what version I'm running there, because
|> 5.0B5 isn't the pre-release for MacOS ;-)
|>
|> So if nothing relating to structures is "improved" before the final
|> release of 5.0, getting a PowerMac may be your only choice.
|>
This is indeed a weird bug - by making a few modifications this runs all the
way to 128 tag names (the stated limit in IDL 4.0.1c) on Digital UNIX. It seems
that the nature of the right hand side of the tag declarations
are important (not the length of the tag name, however..).
Stein Vidar
pro fail
; preparations
ii = 19
b = bytarr(ii)
i = intarr(ii)
l = lonarr(ii)
f = fltarr(ii)
d = dblarr(ii)
s = strarr(ii)
c = complexarr(ii)
x = dcomplexarr(ii)
tyyp=["b","i","l","f","d","s","c","x"]
trun=0
; RSI claims it works up to 128 in IDL 4.01
; on IDL 5.0 the limit should be higher I heard
for try=1,600 do begin
str= ""
; for the try build strings...
for te=1,try do begin
; commata in the middle
if te gt 1 then str=str+","
; build tagname
str=str+"TN_"+strcompress(string(te),/remove_all)
; change tyype every time!
str=str+":"+string(tyyp(trun))
trun=trun+1
if trun eq n_elements(tyyp) then trun=0
endfor
; clear target
target=byte(0)
; build execute-string
estring='target={'+str+'}'
; the big momment: Will it work????
if execute(estring) then begin
print,' Execute was OK with ',te,'Tagnames... length=',strlen(estring)
endif else begin
goto,failing
endelse
endfor
print,'Programm will never come to this point! '
stop
failing:
print, ' Execute failed with ',te,'Tagnames!!! '
end
|
|
|
|
| Re: RSI and the 24. Tagname... [message #8631 is a reply to message #8625] |
Sun, 23 March 1997 00:00 |
gurman
Messages: 82
Registered: August 1992
|
Member |
|
|
In article <5h0mtl$96n@jungle.deceiver.org>, alpha@jungle.deceiver.org
(HSR Sys-Admin) wrote:
> This is my problem since December 1996! RSI was not able to fix the problem
> yet.,,, Another person here in the net was complaining about the same problem
> some weeks ago here...
>
>
> failing systems:
>
> IDL 4.01A on VMS
> IDL 4.01B on VMS
> IDL 4.01A on DU 3.2C
> IDL 4.01A on DU 3.2G
> IDL 4.01B on DU 3.2C
> IDL 4.01B on DU 3.2G
>
> for IDL 5.0b3 on VMS (cannot give you the result due to contract)
>
> on every platform the programm fails at 24. Tagname!
>
> they advised me to use an iterative create_struct fro it, but this can due to
> the huge variety of array form a huge programm to simulate. A simple
simulation
> in our programm now give us only 1% of the speed we expected.
> ------------------------------------------------------------ ------------------
Herr Dr. Roepcke -
Under 5.0b5, about which I guess it's Ok to speak in public because
it;s a _public_ pre-release, I get:
.
.
.
Execute was OK with 37 Tagnames...
Execute was OK with 38 Tagnames...
& END
^
% Program code area full.
Execute failed with 39Tagnames!!!
under OpenVMS 7.1, and somewhat more spectacularly with 32 tagnames under
DU4.0B.
Interestingly, on a PowerMacintosh running MacOS 7.6, I get:
Execute failed with 51Tagnames!!!
but I don't think I can tell you what version I'm running there, because
5.0B5 isn't the pre-release for MacOS ;-)
So if nothing relating to structures is "improved" before the final
release of 5.0, getting a PowerMac may be your only choice.
By the way "tstr" must be changed to "str to make your codelet work.
Best wishes,
Joe Gurman
--
Joseph B. Gurman / NASA Goddard Space Flight Center/ Solar Data Analysis Center / Code 682 / Greenbelt MD 20771 USA / gurman@gsfc.nasa.gov / gurman@ari.net
| Federal employees are still prohibited from holding opinions while at work. Any opinions expressed herein must therefore be someone else's. |
|
|
|
|
| Re: RSI and the 24. Tagname... [message #8632 is a reply to message #8631] |
Sun, 23 March 1997 00:00 |
Harald Frey
Messages: 41
Registered: March 1997
|
Member |
|
|
HSR Sys-Admin wrote:
>
> This is my problem since December 1996! RSI was not able to fix the problem
> yet.,,, Another person here in the net was complaining about the same problem
> some weeks ago here...
>
> failing systems:
>
> IDL 4.01A on VMS
> IDL 4.01B on VMS
> IDL 4.01A on DU 3.2C
> IDL 4.01A on DU 3.2G
> IDL 4.01B on DU 3.2C
> IDL 4.01B on DU 3.2G
>
> for IDL 5.0b3 on VMS (cannot give you the result due to contract)
>
> on every platform the programm fails at 24. Tagname!
>
I tried the program under IDL 4.0.1c on a SUN SPARCstation 5 and got the
result that it was OK until 45 Tagnames
Execute was OK with 44 Tagnames...
Execute was OK with 45 Tagnames...
& END
^
% Program code area full.
Execute failed with 46Tagnames!!!
Harald Frey
Space Sciences Laboratory
University of California at Berkeley
hfrey@ssl.berkeley.edu
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
