KNOWN BUG WITH EXCEPTION PAGES?? - Websphere

This is a discussion on KNOWN BUG WITH EXCEPTION PAGES?? - Websphere ; I and my fellow Java programmer have, over the months, discovered that the particular copy of WebSphere our organization has has many known bugs. So far, we have been able to find fixes for these somewhere on the web, and ...

+ Reply to Thread
Results 1 to 7 of 7

Thread: KNOWN BUG WITH EXCEPTION PAGES??

  1. KNOWN BUG WITH EXCEPTION PAGES??

    I and my fellow Java programmer have, over the months, discovered that the particular copy of WebSphere our organization has has many known bugs. So far, we have been able to find fixes for these somewhere on the web, and implemented the fixes (we don't have the authority to upgrade the co's testing or production versions). But there is still strange behavior--strange in that it doesn't happen with the same code run in other containers, like Tomcat or SAS.

    Does anyone know of a bug or a bug fix for the following problem, which I have only ever seen on WAS:

    I've copied verbatim(as a last resort) source code to create our own custom exception-handling page. One of these is necessary both because it shields the user from meaningless stack traces and ugly diagnostic info, and because if the exception thrown was of a certain custom type (class), we need to display specific content for that case. The source code is good: it comes from an IBM source, and it works normally on other containers--just not on WAS. 60% of the time when an exception is thrown (for example, deliberately using a test page, or naturally) the browser window displays our custom Exception page, and properly reads the implicit "exception" object, programmatically displaying its type, message, and other details.

    However, the other 40% of the time, our custom exception page STILL DISPLAYS with the right content---BUT (this is reeeeeeally bizarre) INCLUDED (INLINED) WITHIN THE SAME PAGE WHICH THREW THE EXCEPTION. FOR EXAMPLE, THE ENTIRE EXCEPTION PAGE MIGHT BE DISPLAYED WITHIN A TABLE CELL () OF A TABLE INSIDE THE ORIGINAL PAGE WHICH THREW IT!!! IT'S HARD TO DESCRIBE IN WORDS, I WILL TRY TO ATTACH A SCREENSHOT SO YOU CAN VISUALIZE WHAT I MEAN.

    It almost seems to be working like the tag! Although no where in either page is that directive used! But it shrinks-down our exception-page's content, and displays it inside of the original throwing page! I've talked to Sun, and numerous other JSP2.0 experts--and none of them have ever seen or heared of anything like that before! At least not with any of the non-IBM containers.

    What the heck is going on? Why is there no reference to any problem like this anywhere in the JSP2.0 literature? I will try to attach some screenshots for you.

    Please don't post a reply unless you have a concrete answer. If you are just speculating out loud, please don't post your speculations until you have a concrete answer. In other words, if the word "maybe" appears anywhere in your answer, don't post it.


    Also, even if you test fire an SQLException or IOException, or any type of exception of any class is thrown--the only class that ever is readable from implicit exception object is class InvocationTargetException (a meaningless wrapper around real exceptions). No matter what the originating exception's class--it only is recognized by WebSphere as an InvocationTargetException (.getClass(), .getMessage(), .toString()...

  2. Re: KNOWN BUG WITH EXCEPTION PAGES??

    I've been working with WebSphere for quite a long time now, and haven't
    managed to reproduce this problem, though I've read other folk having
    strange experiences with exception pages in WAS.

    What version of WAS are you using (WAS_HOME/bin/versionInfo) ?
    Any way you can post the code or fragments of it? (Or just the particular
    IBM page reference)
    Could you include the web.xml file for the web module please?

    Peter Bennett
    Distributed Systems Professional Services Ltd. (Maybe 8-)

    wrote in message
    news:404161764.1184169727186.JavaMail.wassrvr@ltsg was010.sby.ibm.com...
    >I and my fellow Java programmer have, over the months, discovered that the
    >particular copy of WebSphere our organization has has many known bugs. So
    >far, we have been able to find fixes for these somewhere on the web, and
    >implemented the fixes (we don't have the authority to upgrade the co's
    >testing or production versions). But there is still strange
    >behavior--strange in that it doesn't happen with the same code run in other
    >containers, like Tomcat or SAS.
    >
    > Does anyone know of a bug or a bug fix for the following problem, which I
    > have only ever seen on WAS:
    >
    > I've copied verbatim(as a last resort) source code to create our own
    > custom exception-handling page. One of these is necessary both because it
    > shields the user from meaningless stack traces and ugly diagnostic info,
    > and because if the exception thrown was of a certain custom type (class),
    > we need to display specific content for that case. The source code is
    > good: it comes from an IBM source, and it works normally on other
    > containers--just not on WAS. 60% of the time when an exception is thrown
    > (for example, deliberately using a test page, or naturally) the browser
    > window displays our custom Exception page, and properly reads the implicit
    > "exception" object, programmatically displaying its type, message, and
    > other details.
    >
    > However, the other 40% of the time, our custom exception page STILL
    > DISPLAYS with the right content---BUT (this is reeeeeeally bizarre)
    > INCLUDED (INLINED) WITHIN THE SAME PAGE WHICH THREW THE EXCEPTION. FOR
    > EXAMPLE, THE ENTIRE EXCEPTION PAGE MIGHT BE DISPLAYED WITHIN A TABLE CELL
    > () OF A TABLE INSIDE THE ORIGINAL PAGE WHICH THREW IT!!! IT'S HARD TO
    > DESCRIBE IN WORDS, I WILL TRY TO ATTACH A SCREENSHOT SO YOU CAN VISUALIZE
    > WHAT I MEAN.
    >
    > It almost seems to be working like the tag! Although no
    > where in either page is that directive used! But it shrinks-down our
    > exception-page's content, and displays it inside of the original throwing
    > page! I've talked to Sun, and numerous other JSP2.0 experts--and none of
    > them have ever seen or heared of anything like that before! At least not
    > with any of the non-IBM containers.
    >
    > What the heck is going on? Why is there no reference to any problem like
    > this anywhere in the JSP2.0 literature? I will try to attach some
    > screenshots for you.
    >
    > Please don't post a reply unless you have a concrete answer. If you are
    > just speculating out loud, please don't post your speculations until you
    > have a concrete answer. In other words, if the word "maybe" appears
    > anywhere in your answer, don't post it.
    >
    >
    > Also, even if you test fire an SQLException or IOException, or any type of
    > exception of any class is thrown--the only class that ever is readable
    > from implicit exception object is class InvocationTargetException (a
    > meaningless wrapper around real exceptions). No matter what the
    > originating exception's class--it only is recognized by WebSphere as an
    > InvocationTargetException (.getClass(), .getMessage(), .toString()...




  3. Re: KNOWN BUG WITH EXCEPTION PAGES??

    Hi,

    First, please read newsgroup etiquette : http://www.newsgroups.com/articles/e...newsgroups.htm.

    Second, when using custom error pages I have never seen this behavior in WebSphere 6.1.x or 6.0.x. The servlet must throw the exception in order for the custom except page be sent back to the user.

    I have seen this behavior when the error is caught and processing still occurs in the servlet, leading to weird looking pages. In this case, this is an application error.

    Brian

  4. Re: KNOWN BUG WITH EXCEPTION PAGES??

    > Please don't post a reply unless you have a concrete answer. If you are
    > just speculating out loud, please don't post your speculations until you
    > have a concrete answer. In other words, if the word "maybe" appears
    > anywhere in your answer, don't post it.


    Who do you think will be willing to help you with this kind of speech ?

    I *speculate* that nobody will reply to you. *Maybe* I'm wrong. But *maybe*
    not...


  5. Re: KNOWN BUG WITH EXCEPTION PAGES??

    Thanks for your reply. In reading your reply, it seems your instance of WebSphere is not behaving the same way ours is (no such bug). You describe the knowns of exception-handling, which we also know and have set up our custom exception JSP accordingly. That is reassuring, because it means the problem we are all experiencing is specific to our (older) installed instance, rather than anything we were doing wrong in our code (which is why no one else has ever experienced this same problem, and no references to it at all in any JSP2.0 documentation. Looks like, until our company updates to a more recent version on its test and production servers.

  6. Re: KNOWN BUG WITH EXCEPTION PAGES??

    And that is exactly the reason why human beings make horrible court witnesses: they imagine and see meaning and emotion in to things which have neither.

    If you actually literally read the words I wrote, you can find no profanity, vulgarity, insulting words, or ugliness. Because it was written text, and you could not hear any 'inflection' in a spoken voice, probably what happened is that you read it, and your mind imagined a person speaking those words with anger in their voice. But that is entirely in your mind. I had no emotion or anger in that text, and I did not feel anger when I wrote it. I was simply making an emotionless request. You have no way of knowing what inflection was intended--and you were unaware that if you heared me speaking those same words, you would have heard them in a flat, calm, matter-of-fact voice.

    The actual content written, itself, does not constitute anything negative, purely neutral and emotionless.

    Here is a good example which illustrates this facet of human beings: I once asked a friend: "So you want me to deploy the application to the server?"

    I asked in a completely matter-of-fact tone, and my intent was simply to confirm that was what he wanted.

    But instead of simply answering "yes"--he got angry and replied "Why not?! What's wrong with that?! Why shouldn't you?"

    In other words, there was a whole imaginary world in his head--which had no basis in reality. He was seeing more emotion into my words than actually were there. I simply was asking for the same reason a computer would ask: to find out the answer, yes or no. I think the same phenomenon of Homo Sapiens happened here. There was no emotion in my voice when I wrote my original post--just like there isn't now. Remember Spock? The character who spoke and acted emotionless in Star Trek? If you remember, the other characters tended to mistake his directness for an emotion, too--he was one of the least popular characters to the other characters, for the same reason.

  7. Re: KNOWN BUG WITH EXCEPTION PAGES??

    Like you said, emotion is easily added where there isn't. However, I use to
    stick to words and to not try to add emotion where there isn't.

    However, your post reads with emotion even if you didn't intend: using a
    "what the heck" expression and writing in capital letters. Add to this that
    you refuse any form of help except *the* solution and pretend to tell people
    how they should speak / write. The first is emotion you add to the post and
    the second creates emotion to the reader.

    Nevertheless, good to see you had a peaceful mind at the time of writing and
    I'll stay to that. Case closed for me.


+ Reply to Thread