jre/lib/ext and classpath

[David Segall]

The relevant sentences from http://mindprod.com/jgloss/jar.html are:

According to Sun's JAWS FAQ, in JDK 1.5,
Java Web Start still does not support Class-Path.
Up until JDK 1.5, java.exe too ignored this entry.


Note this bug report, which sun closed, as evidence I did not make
this up.

http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=4356783

The summary of this bug report and almost all the posts refer to the
fact that _only_ Class-Path is used, not that Class-Path is ignored.

A post from JMS reports a bug which supports the above sentence but
which, if true, should be a completely separate bug. It is confusing
because JMS says he was wrong in a post that precedes his bug report.

 

[Raymond DeCampo]

FUD mean Fear Uncertainty and Doubt. A marketing tactic to freeze the
buyer from making a decision by leaving an impression that a better
product might be available imminently, and that there are mysterious
fatal flaws with the existing ones.

Saying something erroneous is not the same thing as spreading FUD.

You've repeatedly claimed that applets cannot access files that are not
"downstream" from the applet codebase, for your definition of the term
"downstream". I've challenged you to provide a SSCCE of this and you
have not.

I consider this spreading FUD.

Ray

 

[Raymond DeCampo]

Please at least quote me if you want to accuse me.

Please be specific with both your charges. I can't very well correct
the entries if I don't know the sentences you are referring to.

The message you have replied to contained a link to the erroneous
information, as well as a description of what was wrong. For some
reason you have trimmed this information from your reply.

Ray

 

[Raymond DeCampo]

The relevant sentences from http://mindprod.com/jgloss/jar.html are:

According to Sun's JAWS FAQ, in JDK 1.5,
Java Web Start still does not support Class-Path.
Up until JDK 1.5, java.exe too ignored this entry.


Note this bug report, which sun closed, as evidence I did not make
this up.

http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=4356783

This bug report does not support your claims. First of all, it is
"closed, will not be fixed" because the software provides "the intended
behavior" "as described in the SDK docs."

Secondly, the workaround listed says to "use the 'Class-Path' header in
the manifest."

Finally, the report is against 1.3 and has nothing to do with 1.5

Ray

 

[Roedy Green]

The message you have replied to contained a link to the erroneous
information, as well as a description of what was wrong. For some
reason you have trimmed this information from your reply.

But you did not check what I said or quote it. What I said was
neither erroneous nor FUD.

You are playing an Eric Berne style game.

 

[Raymond DeCampo]

But you did not check what I said or quote it. What I said was
neither erroneous nor FUD.

You are playing an Eric Berne style game.

I'm not playing a game, I'm trying to get you to fix bad information on
your website.

The text has been extensively quoted in this thread, but if you insist,
I will quote it again here:

According to Sun's JAWS FAQ, in JDK 1.5, Java Web Start still
does not support Class-Path. Up until JDK 1.5, java.exe too
ignored this entry.

- from http://mindprod.com/jgloss/jar.html

The erroneous statement is the second sentence. (I include the first
only because the second is incomprehensible without it. I was able to
substantiate it at
http://java.sun.com/j2se/1.5.0/docs/guide/javaws/developersguide/faq.html#209.
I think including this link in your page would be an improvement.)

Roedy, I seemed to have touched a nerve by accusing you of spreading
FUD, so I apologize. I am one of those sarcastic, caustic
northeasterners (in the US) and sometimes by voice does not carry well
in the usenet. I know that you are not intentionally trying to spread
incorrect information. My intention was not to antagonize you and start
a pointless flamewar. I will endeavor to be more diplomatic in the future.

These are the facts concerning the Class-Path entry in manifest files of
jars:

- When started via the -jar option, the JRE will use the contents of
the Class-Path entry as the classpath for the JVM. I have seen this
work for 1.2, 1.3 and 1.4. There is no reason to expect it does not
work for 1.1 also.
- When a jar is included in the classpath, its Class-Path entry is
included in the classpath for the JVM. (This allows you to ship a
library consisting of multiple jars and have your clients include only
the one jar in the classpath.) I have seen this work in 1.4 and 1.5.
- I am not aware of any bugs affecting this feature, although there may
be some. But there is certainly not one that prevents it from working
for all JVMs prior to 1.5.
- Bug 4356783
(http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=4356783) in
particular, which you have cited in other posts to support the above
statement, is a) not a bug, b) recommends that the Class-Path manifest
entry be used when starting the JVM with the -jar option.

Ray

Ray

 

[Roedy Green]

Roedy, I seemed to have touched a nerve by accusing you of spreading
FUD, so I apologize. I am one of those sarcastic, caustic
northeasterners (in the US) and sometimes by voice does not carry well
in the usenet. I know that you are not intentionally trying to spread
incorrect information. My intention was not to antagonize you and start
a pointless flamewar. I will endeavor to be more diplomatic in the future.

Thank you. I will take your word for it, and put it at low priority on
my to do list to verify.

I had not done experiments. So when this sort of thing happens and
person A says X and person B says Y, I tend to believe the person who
seems least likely out for my scalp. You seemed primarily out for my
scalp and at least in the early stages it looked as though your
objection was purely based on that I, a habitual fudster, had said it,
not on any practical experiments with Class-Path.

Yes, you really hit a nerve. You have no idea how many thousands of
man hours I have devoted trying to make the Java glossary useful,
which in turn promotes a platform-independent language, accessible
even to people in the third world. It is about 200 MB. That is one
heck of a lot of keystrokes for one person to pump out. It does not
generate any revenue other than the Google ads to pay the website bill
(but not my Internet access). Of course there are errors in it. I
spend hours every day working to get rid of them. When you accused me
of deliberately trying to sabotage Java I wanted you wring your neck.

Even Sun, in a minor way, recognised my effort by giving me a free
Java jacket. They would not do that to someone they considered an
enemy.

 

[Raymond DeCampo]

Thank you. I will take your word for it, and put it at low priority on
my to do list to verify.

Fair enough.

I had not done experiments. So when this sort of thing happens and
person A says X and person B says Y, I tend to believe the person who
seems least likely out for my scalp. You seemed primarily out for my
scalp and at least in the early stages it looked as though your
objection was purely based on that I, a habitual fudster, had said it,
not on any practical experiments with Class-Path.

Yes, you really hit a nerve. You have no idea how many thousands of
man hours I have devoted trying to make the Java glossary useful,
which in turn promotes a platform-independent language, accessible
even to people in the third world. It is about 200 MB. That is one
heck of a lot of keystrokes for one person to pump out. It does not
generate any revenue other than the Google ads to pay the website bill
(but not my Internet access). Of course there are errors in it. I
spend hours every day working to get rid of them. When you accused me
of deliberately trying to sabotage Java I wanted you wring your neck.

I never meant to imply you were an enemy of Java. In my mind, FUD
doesn't necessarily imply that. You've already explained what that
phrase means to you. Since we had different meanings, I retracted my
statement. I try to live my life by the credo correcting mistakes
rather than covering and compounding them. I certainly made a mistake
in the way I approached you on this matter and I appreciate the
opportunity to have a second chance.

I also recognize that you have made huge contributions to the community.
More than I likely ever will.

Even Sun, in a minor way, recognised my effort by giving me a free
Java jacket. They would not do that to someone they considered an
enemy.

Ray