[ nice-Bugs-798317 ] linker problem?

Bugs item #798317, was opened at 2003-09-01 02:06
Message generated for change (Comment added) made by bonniot
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=112788&aid=798317&group_id=12788

Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Submitted By: Isaac Gouy (igouy)
Assigned to: Nobody/Anonymous (nobody)
Summary: linker problem?

Initial Comment:
Refer to SieveBeginnerExample.

I moved the 'var count' declaration inside the loop like 

this:



   let stop = 8192;

   let isPrime = new boolean[stop+1];



   while (n-- > 0){ 

      var count = 0;



and this happened:







I:\pls\Nice\bench 2\sieve>nicec --sourcepath=.. -d=. -a 

sieve.jar sieve

nice.lang: parsing

sieve: parsing

sieve: typechecking

sieve: generating code

sieve: linking

<unknown>:0: Literals: Internal error:java.lang.Error: 

gnu.expr.PrimProcedure does not implement 

Externalizable

java.lang.Error: gnu.expr.PrimProcedure does not 

implement Externalizable

        at gnu.expr.LitTable.error(LitTable.java:72)

        at gnu.expr.LitTable.writeObject(LitTable.java:207)

        at gnu.expr.LitTable.emit(LitTable.java:35)

        at gnu.expr.Literal.emit(Literal.java:118)

        at gnu.expr.LambdaExp.generateClassInit

(LambdaExp.java:1125)

        at gnu.expr.Compilation.compileAll

(Compilation.java:1604)

        at gnu.expr.Compilation.generate

(Compilation.java:1571)

        at gnu.expr.Compilation.compile

(Compilation.java:1539)

        at gnu.expr.Package.compileToFiles

(Package.java:100)

        at bossa.modules.Package.finishCompilation

(Package.java:362)

        at bossa.modules.Package.link(Package.java:341)

        at mlsub.compilation.fun.lambda23

(~/Nice/src/mlsub/compilation:47)

        at mlsub.compilation.fun.apply1

(~/Nice/src/mlsub/compilation)

        at gnu.expr.ModuleMethod.apply1

(ModuleMethod.java:89)

        at nice.lang.fun.foreach

(~/Nice/stdlib/nice/lang:116)

        at nice.lang.dispatch.foreach

(~/Nice/classes/nice/lang)

        at mlsub.compilation.fun.compileComponent

(~/Nice/src/mlsub/compilation:47)

        at mlsub.compilation.fun$make.lambda16(Unknown 

Source)

        at mlsub.compilation.fun$make.apply1(Unknown 

Source)

        at gnu.expr.ModuleMethod.apply1

(ModuleMethod.java:89)

        at nice.lang.fun.foreach

(~/Nice/stdlib/nice/lang:116)

        at nice.lang.dispatch.foreach

(~/Nice/classes/nice/lang)

        at mlsub.compilation.fun.make

(~/Nice/src/mlsub/compilation:60)

        at nice.tools.compiler.fun.compile

(~/Nice/src/nice/tools/compiler:40)

        at nice.tools.compiler.console.fun.compile

(~/Nice/src/nice/tools/compiler/console:168)

        at nice.tools.compiler.console.fun.main

(~/Nice/src/nice/tools/compiler/console:184)

sieve: writing in archive

nice.lang: writing in archive

----------------------------------------------------------------------

>Comment By: Daniel Bonniot (bonniot)
Date: 2003-09-01 20:15

Message:
Logged In: YES 
user_id=88952

Yes, it can be surprising to have the power of first-class

functions ;-)



That said, I agree this is a bit problematic. Not tragic,

because more or less the only thing you can that will not

find out the type mismatch is toString, and you will the

result is strange... Still, I think the solution in the long

term is to restrict the type of toString to "printable"

types (the ones that define a meaningful toString method).

This needs some improvements in the type system, so it will

take some time...



----------------------------------------------------------------------

Comment By: Isaac Gouy (igouy)
Date: 2003-09-01 16:12

Message:
Logged In: YES 
user_id=536291

> Could you please send a self-contained bug report?

Sorry



>count method ... the code is valid

Gosh!

----------------------------------------------------------------------

Comment By: Daniel Bonniot (bonniot)
Date: 2003-09-01 13:00

Message:
Logged In: YES 
user_id=88952

The local variable count is not visible at the println, but

there is a count method defined in nice.lang collections,

which is the one is refered to. Since anything can be

converted to a String (even methods), the code is valid.

----------------------------------------------------------------------

Comment By: Arjan Boeijink (arjanb)
Date: 2003-09-01 12:46

Message:
Logged In: YES 
user_id=688815

Simplified testcase:



void main(String[] args){

   var n = 1;

   while (n-- > 0){ 

      var count = 0;

   }

   println("Count: " + count); 

}



This code should give a user error at checking but doesn't.

----------------------------------------------------------------------

Comment By: Daniel Bonniot (bonniot)
Date: 2003-09-01 10:19

Message:
Logged In: YES 
user_id=88952

Could you please send a self-contained bug report?

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=112788&aid=798317&group_id=12788


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf