ant-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Magesh Umasankar" <umag...@apache.org>
Subject Re: ExecuteJava and initializeClass
Date Thu, 17 Jan 2002 21:19:36 GMT
Here is a recap of what I have tested and confirmed:

With the reflection based patch applied, I performed 
the test laid out at 
http://nagoya.apache.org/bugzilla/show_bug.cgi?id=1362
on the following platforms/JDKs with these results:

WindowsNT, 98 
===========
Sun JDKs 1.1.6, 1.1.7A, 1.1.8, 1.2.1, 1.2.2, 1.3.0, 1.3.1
IBM JDK 1.3
Result: B CLASS INITIALIZATION gets printed.

IBM JDK 1.1.8
Result: B CLASS INITIALIZATION does *not*
get printed.  But this is not really an Ant issue because
it does not get printed even when java is invoked from the 
command line.

FreeBSD:
======
FeeBSD port of JDK 1.1.8:
Result: B CLASS INITIALIZATION gets printed.

Platforms/JDKs *not* tested:
==================
Linux - Blackdown/IBM/SUN

As I mentioned earlier, this patch will force the static
initializers to be invoked without having to create a 
new instance of the object.  

The existing code works on all environments mentioned
above, including IBM JDK 1.1.8, but it carries the
undesirable side effect of calling the constructor.
http://nagoya.apache.org/bugzilla/show_bug.cgi?id=4107

I think it is reasonable to say that Ant's <java> task
will work just like that of the jdk's - no more, no less.
Do you agree?  If so, I will commit this patch as my tests
have given me a reasonable level of confidence - of course
it hasn't been tested on linux yet ;-(

Let me also know if I need to be testing against some
other earlier versions of the JDK - Icouldn't lay my hands
on Sun'c JDKs earlier than 1.1.6...

Thanks,
Magesh



--
To unsubscribe, e-mail:   <mailto:ant-dev-unsubscribe@jakarta.apache.org>
For additional commands, e-mail: <mailto:ant-dev-help@jakarta.apache.org>


Mime
View raw message