Re: Bogus NullPointerExceptions
Pay attention and read the entire message before you start steaming,
there is more than just vivid rudeness in this reply. There is also an
actual attempt at providing some help. Only Allah knows why... (LOL).
But that all depends on whether you want some help or are just out to
assert yourself.
Twisted wrote:
Tom Forsmo wrote:
Twisted wrote:
I suppose every newcomer to this thread will need this explained
separately and again.
Don't be condescending. You have a bad design and you insist on fixing
it by adding more complexity, synchronisation.
Don't be insulting. There is nothing wrong with my design.
What insult did I commit? Is it that asked you not to be
condescending? Oh no! some little person asking you to behave like a
grown up, such insolence! ... Ahh, now I get it. Its because I said your
design was bad, now that's a real insult. You should bang my head against
the wall until it pops. Or better yet, you should just raise your voice
so everybody can hear the mighty Twisted speak.
You need to deflate quite a bit. This is a public forum, where the aim
is to try to provide help, You have already shown us, the last couple of
weeks, that you are not quite the genius you envision yourself to be.
There are messages scattered around from the last couple of weeks, where
you seem to have some problem with your program and the fault is
definitely not yours, it has to be java, it has to be...
This leads me and others to notice that maybe you should be looking at
the code you write.
I'd like to
see you try to suggest a better way to:
Because there can only be one solution for this task? what arrogance!
* Delete a file in a particular subtree and then
* If the directory it's in is now empty, delete that, and if that
leaves its parent directory entry, delete that too, and so forth, while
remaining confined to that subtree (whose root, even if it becomes
empty, isn't deleted).
Did you ever read ESRs "How to ask questions the right way" (I'll leave
it to you to find it on the net). You should have noticed that, in at
least one of the replies to your post, it was suggested that you should
try to look for the real problem, instead of creating a quick fix. That
should suggest to you that 1) maybe the problem lies deeper than what it
seems initially 2) maybe your design is not the best way to solve this
particular problem 3) to solve this problem you should look at it in a
larger perspective or from a different angle. 4) Maybe if you explain
more than just a little bit of the problem you can get some real answers.
That's why its not over, you may ignore any further comments if you
wish, but you should not tell others what to do with the thread in a
public forum.
You seem to misunderstand. Everything has already been said, and now
you are just going around in circles
No, I have not misunderstood anything, I have read all the posts and I
know what the posters say. But you do not have time to listen, as you
throw out insults instead on focusing on what the posters are trying to
say to you.
(and forcing me to do so as well,
to rebut insulting BS like what you just posted) saying the same things
repeatedly.
some more hot air...
You are awfully free with criticism, but seem to lack any
constructive suggestions -- or if not, you're keeping them to yourself.
I suggest you reverse that pattern -- keep your criticisms to yourself
and make any constructive suggestions you may have public. (Yes, that
does mean that in the event you have no constructive suggestions you
should simply shut up.)
Fortunately you do not have the possibility to control what I say, when
I say it and where I say it.
You don't seem to have noticed that people here have actually tried to
help you, but you have been so busy cutting them of with rude remarks
and self assertions that they simply cant be bothered to help you any more.
But I suspect both of us have better things to do than to continue this
pointless debate. Obviously you disagree with me; equally obviously
neither of us is likely to change the other's mind. Slinging insults
around won't do anything but waste your time slinging them and mine
cleaning up after you. So let's both just go home.
You might have been out flying high, but I have been home all the time
with my feet planted firmly on the ground.
Now, if you are ready to listen, I will give you the suggestions you
think I am unable to provide.
But first, there are some things you need to answer to get proper help,
because as several people have already stated your design has a bad
smell and should probably be redesigned.
- what is the purpose of what you are trying to do?
- why do you need a deep directory structure with files scattered in it?
- why do you need support for multi threaded file operations within the
same directory structure?
- can other solutions help you better?
-
the main criticisms are:
- the design and implementation you have chosen is not very maintainable
- the code its more complicated than it needs to be
- i.e. there are too many difficult to control side effects in the
code.
which both can lead to more headache in the future.
here are some suggestions:
1- don't use multi threading
- but that only solves your immediate problem, so you could almost
just continue to use multi threading.
2- add some proper file error handling code, to allow you to recover
from file errors, as the file system is not very transaction based.
3- not use deep directory structure but some other method instead, such
as dotted hierarchical file names in one directory.
4- design the directory operations in a singleton object in its own
thread, with a message queue to send commands through.
- at least now we are getting closer to some good design for reuse
and maintainability.
5- a recursive clean up method instead of the iterative version you
have chosen.
6- extend 4 with a separate TimerTask thread run every x seconds.
So the only thing you main thread, or x number of threads,
does, is to delete a single file. The clean up thread will take care
of the rest for you. This single file would then only be named in one
thread so there will be no problems. And the operation that creates a
file in the directory structure must also be able to create any
missing directories, so it should be able to handle disappearing
directories
7- or even better, combine 4,5 and 6 into one solution, where the
TimerTask is to add a message in the message queue signalling clean-up.
tom