www.digitalmars.com         C & C++   DMDScript  

digitalmars.D.bugs - $ should be a compile-time constant, if used inside a const array

reply Don Clugston <dac nospam.com.au> writes:
----
const char [] elephant = "abcdef"[2..$];
----
newbug.d(1): non-constant expression "abcdef"[2..cast(int)(__dollar)]

Likewise for:
----------
const char giraffe = "abc"[$];
------------

inside a subscript expression.

dmd -c -o- bug.d
(has anyone used -o- before?)
Dec 07 2005
next sibling parent reply Oskar Linde <oskar.lindeREM OVEgmail.com> writes:
Don Clugston wrote:

 ----
 const char [] elephant = "abcdef"[2..$];
 ----
 newbug.d(1): non-constant expression "abcdef"[2..cast(int)(__dollar)]
 
 Likewise for:
 ----------
 const char giraffe = "abc"[$];
 ------------

 inside a subscript expression.


Useful for obfuscated code...:

class MyArrayContainer {
         int length() { return 4; }
         int opIndex(int i) { return i; }
}

void main() {
         MyArrayContainer arr = new MyArrayContainer;

         int __dollar = 7;
         writef("%d ",arr[$-1]);
         arr[$=15];
         writef("%d\n",arr[++$]);
}

prints 6 16

:)



 dmd -c -o- bug.d
 (has anyone used -o- before?)


Hmm... -c -o- seems like it might as well be a NOP...?

/Oskar
Dec 07 2005
parent Don Clugston <dac nospam.com.au> writes:
Oskar Linde wrote:

 Don Clugston wrote:
 

 ----
 const char [] elephant = "abcdef"[2..$];
 ----
 newbug.d(1): non-constant expression "abcdef"[2..cast(int)(__dollar)]

 Likewise for:
 ----------
 const char giraffe = "abc"[$];
 ------------

 length, inside a subscript expression.


 
 
 Useful for obfuscated code...:
 
 class MyArrayContainer {
         int length() { return 4; }
         int opIndex(int i) { return i; }
 }
 
 void main() {
         MyArrayContainer arr = new MyArrayContainer;
 
         int __dollar = 7;
         writef("%d ",arr[$-1]);
         arr[$=15];
         writef("%d\n",arr[++$]);
 }
 
 prints 6 16
 
 :)


Cool! (If you like that sort of thing)
Actually, D has some great obfuscation potential.
As discussed below, you can print output at compile time, so the top 
half of a file could be a compile-time program, and
the bottom half could be complete garbage that won't even compile.
:-)



 dmd -c -o- bug.d
 (has anyone used -o- before?)


 
 
 Hmm... -c -o- seems like it might as well be a NOP...?


Actually, it isn't.  I made a batch file, dd.bat, consisting of just
 dmd -c -o- %1

Then, using compile-time metafunctions, and using pragma(msg) to print, 
you can treat it as a complete (interpreted) programming language.
eg
dd beer.d
prints the "99 bottles of beer" song. There's no hint anywhere that 
you're using a language which is normally compiled :-).

I use dd when writing metafunctions, to be certain that everything is 
happening at compile time.
Dec 07 2005
prev sibling parent Thomas Kuehne <thomas-dloop kuehne.cn> writes:
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Don Clugston schrieb am 2005-12-07:

 ----
 const char [] elephant = "abcdef"[2..$];
 ----
 newbug.d(1): non-constant expression "abcdef"[2..cast(int)(__dollar)]

 Likewise for:
 ----------
 const char giraffe = "abc"[$];
 ------------

 inside a subscript expression.

 dmd -c -o- bug.d
 (has anyone used -o- before?)


Added to DStress as
http://dstress.kuehne.cn/run/l/length_08_A.d
http://dstress.kuehne.cn/run/l/length_08_B.d
http://dstress.kuehne.cn/run/l/length_08_C.d
http://dstress.kuehne.cn/run/l/length_08_D.d
http://dstress.kuehne.cn/run/l/length_08_E.d

Thomas


-----BEGIN PGP SIGNATURE-----

iD8DBQFDvQ9/3w+/yD4P9tIRAj3vAJ4vPHCWDesCbcetQ7o4nwZI0SpicACfaApa
s2beYhEXNpKb8L5WMcVkwd8=
=wQ1h
-----END PGP SIGNATURE-----
Jan 05 2006