Attachment 1706 Details for Bug 21661 – e-mail to jdt-core-dev@eclipse.org describing the phenomenon

e-mail to jdt-core-dev@eclipse.org describing the phenomenon

e-mail.txt (text/plain), 3.40 KB, created by Robert Klemme

on 2002-07-17 13:01:49 EDT

(hide)

Description:

Filename:

MIME Type:

Creator: Robert Klemme

Created: 2002-07-17 13:01:49 EDT

Size: 3.40 KB

patch

obsolete

>
>hi there!
>
>> Would be interesting to see the actual test case. The compiler may indeed
>> require classfiles indirectly referenced by others you explicitly
>> reference.
>> But if so, I am wondering why we would need more than actually needed to
>> run (could be wrong, but would need the actual test case to assess).
>
>think of this scenario:
>
>public class X {
>public void foo() {
>  if ( getCriterion() ) {
>    Y y = new Y();
>    y.someMethod();
>  }
>}
>
>}
>
>now, as long as getCriterion() evaluates to "false" Y is never loaded by the
>JVM when this is the only place referencing it (or all other places are not
>accessed either).  the jvm spec reads in section 2.17.3 Linking:
>Verification, Preparation, and Resolution:
>
>The Java programming language allows an implementation flexibility as to
>when linking activities (and, because of recursion, loading) take place,
>provided that the semantics of the language are respected, that a class or interface
>is completely verified and prepared before it is initialized, and that
>errors detected during linkage are thrown at a point in the program where some
>action is taken by the program that might require linkage to the class or
>interface involved in the error.
>
>see
>http://java.sun.com/docs/books/vmspec/2nd-edition/html/Concepts.doc.html#22574
>
>the language spec reads in 12.4.1 When Initialization Occurs:
>
>A class or interface type T will be initialized immediately before the first
>occurrence of any one of the following:
>
> - T is a class and an instance of T is created.
>
> - T is a class and a static method declared by T is invoked.
>
> - A static field declared by T is assigned.
>
> - A static field declared by T is used and the reference to the field is
>not a compile-time constant (§15.28). References to compile-time constants must
>be resolved at compile time to a copy of the compile-time constant value, so
>uses of such a field never cause initialization.
>
>see
>http://java.sun.com/docs/books/jls/second_edition/html/execution.doc.html#57946
>
>in other words: if a class is missing, the error is only reported when it is
>actually needed during execution, regardless which policy the vm employs to
>load classes.
>
>> When you say it compiled fine using Ant compile, which compiler did you
>> use in there ?
>
>the one coming with jdk build 1.3.1_04-b02.
>
>> I believe there is little you can do, but trying to add the missing types
>
>thanks to dave i could do this.  i didn't realize the classes are there.
>however, i think this is a more general problem since the antlr people could
>have choosen not to disclose these classes without harm to the correct
>functioning of their library.
>
>my personal preference would be to have a compiler switch for this with the
>same three options "ignore", "warning" and "error" like the others.  the
>semantics would apply to classes referenced from classes that are outside the
>scope of the current project's compilation, typically in jars and zips, but it
>could be a class folder as well.
>
>> (or removing the reference path that leads to them - may be even harder).
>
>yes, this one is not feasible.
>
>> We might want to figure exactly what is going on here, and investigate a
>> fix if proven to be a bug.
>>
>> Could you please enter a bug report against JDT/Core with steps to
>> reproduce ?
>
>ok.  i'll try to set up a minimal scenario that creates the problem.
>
>thank's for listening.
>
>regards
>
>        robert
>

Actions: View

Attachments on bug 21661: 1706 | 1718