Michael Macaluso wrote:
>True. The only thing that leads me to believe it is in the VM is that
>different VM versions react differently, but that may be circumstantial.
My first thought was "Huh? Who said different JVM versions react
differently?", so I went through the message thread and only came up with
these:
<http://lists.apple.com/archives/java-dev/2006/Aug/msg00325.html>
"I am using latest version of system (java 1.5) on a macbook pro"
<http://lists.apple.com/archives/java-dev/2006/Aug/msg00362.html>
No Java version cited.
Then I went back and carefull reread your original reply:
<http://lists.apple.com/archives/java-dev/2006/Aug/msg00366.html>
I suspect you (Michael) are talking about a different bug or bugs than the
original poster (Francois-Paul) presented.
As I read Francois-Paul's original post, it was specifically about a
problem where 2 composed-accent e-acutes in a row works, but 3 or 4 in a
row does not. Your complaint seems to be more general: that some
compositions simply don't work in Java. What's unclear is whether or not
your failing compositions will work or fail in non-Java programs.
The evidence I found is that Francois-Paul's original problem with 2 vs. 3
or more composed e-acutes in a row will occur even when Java is not
involved. Specifically, the problem appears in Apache on 10.4.7 serving
files with 2, 3, or 4 e-actues in a row. I suspect it appears on earlier
OS versions too, but have not confirmed this.
It may be that there's one underlying bug that's causing all these
problems, for both you and Francois-Paul, in both Java and Apache, but I've
seen no evidence of that. If Francois-Paul's bug were causing your
problems, then I'd expect it to appear consistently in J2SE 1.3, 1.4, and
1.5. Yet you say that the problem differs across Java versions, so I don't
see how it can be just one bug.
At this point, I suspect there is at least one bug in an Apple library or
framework that is causing Francois-Paul's problem with 3 or more composed
e-acute's in a row. The evidence is that the problem occurs when only
Apache is used. Since Java isn't involved at all, it's illogical to
conclude that this particular problem lies in the JVM code.
It's possible there are additional bugs that are specific to Java, or
specific to versions of Java, but without a test to isolate them by Java
version, framework, library, etc. there's not much more to say.
At this point, I think the next step is to collect additional evidence.
Create tests for the 2/3 e-acutes problem in a few different programming
languages, or using different libraries, and correlate the results. This
may provide evidence of a more general bug with composed accents, or it may
not. If not, the original bug is still a definite bug, no matter where its
cause actually lies.
-- GG
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Java-dev mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/java-dev/email@hidden
This email sent to email@hidden
