| Re: array subscript conversion [message #55031] |
Wed, 25 July 2007 11:26 |
Foldy Lajos
Messages: 268
Registered: October 2001
|
Senior Member |
|
|
On Wed, 25 Jul 2007, Dick Jackson wrote:
> Hi,
>
> "FÖLDY Lajos" <foldy@rmki.kfki.hu> wrote in message
> news:Pine.LNX.4.64.0707251721430.28405@bifur.rmki.kfki.hu...
>> Hi guys,
>>
>> according to the manual, array subscripts are converted to long (or long64 on
>> 64 bit systems) before use if necessary, so an explicit conversion should not
>> affect the result.
>>
>> IDL> print, !version
>> { x86 linux unix linux 6.3 Mar 23 2006 32 64}
>> IDL>
>> IDL> a=lindgen(10)
>> IDL> print, a[[long(-1ull)]]
>> 0
>> IDL> print, a[[-1ull]]
>> 9
>>
>> Is it a bug or I am missing something?
>
> I think you're expecting -1ull to be negative, but the 'u' in 'ull' means
> 'unsigned'. What you end up with instead of -1 is the largest 64-bit integer
> (this is a nice shortcut when it's actually what you want to do!):
>
No, I have expected LONG(-1ull) to be used for subscripting, which is
really negative. As Mike wrote, integers are not converted to LONG, that
was my wrong assumption.
I am trying to find differences between IDL and FL behavior, and this is
one example. In FL, I always convert non-LONG subscripts to LONG. (It's a
pity we have no formal definition of IDL syntax and semantics.)
regards,
lajos
|
|
|
|
| Re: array subscript conversion [message #55032 is a reply to message #55031] |
Wed, 25 July 2007 11:04  |
Foldy Lajos
Messages: 268
Registered: October 2001
|
Senior Member |
|
|
On Wed, 25 Jul 2007, mgalloy@gmail.com wrote:
> On Jul 25, 9:36 am, FÖLDY Lajos <fo...@rmki.kfki.hu> wrote:
>> Hi guys,
>>
>> according to the manual, array subscripts are converted to long (or long64
>> on 64 bit systems) before use if necessary, so an explicit conversion
>> should not affect the result.
>>
>> IDL> print, !version
>> { x86 linux unix linux 6.3 Mar 23 2006 32 64}
>> IDL>
>> IDL> a=lindgen(10)
>> IDL> print, a[[long(-1ull)]]
>> 0
>> IDL> print, a[[-1ull]]
>> 9
>>
>> Is it a bug or I am missing something?
>>
>> regards,
>> lajos
>
> All I see in the manual about converting subscripts to longs is:
>
> "Subscripts can be any type of vector or scalar expression. If a
> subscript expression is not integer, a longword integer copy is made
> and used to evaluate the subscript."
>
> around the middle of this page:
>
> http://idlastro.gsfc.nasa.gov/idl_html_help/Understanding_Ar ray_Subscripts.html
>
> By "integer", I think they mean the more general any integer type:
> byte, integer, long, etc. And in your case, the type is "integer", so
> no conversion is made.
>
> I am I missing another statement in the help that says something more
> explicit?
>
> Mike
> --
> www.michaelgalloy.com
>
You are right. I remembered "not long" instead of "not integer". And also
remembered IDL_MEMINT from idl_export.h, which represents memory offsets
and sizes, and assumed that IDL always uses IDL_MEMINT for subscripting
internally.
Second try:
IDL> a=lindgen(10)
IDL>
IDL> print, a[4294967296ll]
0
IDL> print, a[[4294967296ll]]
9
In the first case the scalar subscript is "integer", so no conversion is
needed. But it is converted to LONG (=0). Why?
regards,
lajos
|
|
|
|
| Re: array subscript conversion [message #55033 is a reply to message #55032] |
Wed, 25 July 2007 10:46 |
Dick Jackson
Messages: 347
Registered: August 1998
|
Senior Member |
|
|
Hi,
"F�LDY Lajos" <foldy@rmki.kfki.hu> wrote in message
news:Pine.LNX.4.64.0707251721430.28405@bifur.rmki.kfki.hu...
> Hi guys,
>
> according to the manual, array subscripts are converted to long (or long64 on
> 64 bit systems) before use if necessary, so an explicit conversion should not
> affect the result.
>
> IDL> print, !version
> { x86 linux unix linux 6.3 Mar 23 2006 32 64}
> IDL>
> IDL> a=lindgen(10)
> IDL> print, a[[long(-1ull)]]
> 0
> IDL> print, a[[-1ull]]
> 9
>
> Is it a bug or I am missing something?
I think you're expecting -1ull to be negative, but the 'u' in 'ull' means
'unsigned'. What you end up with instead of -1 is the largest 64-bit integer
(this is a nice shortcut when it's actually what you want to do!):
IDL> help,-1ull
<Expression> ULONG64 = 18446744073709551615
That's why a[[-1ull]] gives 9, as the subscript is (somewhat) larger than the
maximum index in the array a. It is similar to a[[11]] below...
IDL> help,long(-1ull)
<Expression> LONG = -1
This is similar to a[[-1]] below...
IDL> print,a[[-1]]
0
IDL> print,a[[11]]
9
And we know this is different from simple subscripting which doesn't allow
out-of-range values:
IDL> print,a[-1]
Attempt to subscript A with <INT ( -1)> is out of range.
Execution halted at: $MAIN$
IDL> print,a[11]
Attempt to subscript A with <INT ( 11)> is out of range.
Execution halted at: $MAIN$
--
Cheers,
-Dick
--
Dick Jackson Software Consulting http://www.d-jackson.com
Victoria, BC, Canada +1-250-220-6117 dick@d-jackson.com
|
|
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
