Re: Exception Misconceptions
"tanix" <tanix@mongo.net>
Ok, the situation is this:
[whole story can be replaced with "you have to keep info on arbitrary number
of incoming network packets for arbitrary long time"]
That much was clear from previous posts. And it has little to do with memory
leaks or their prevention.
I see no problem to create your thing without leaks, as for every packet
there is a clear set of conditions when to keep and when to let go.
OTOH your requirements provide no upper limit resource usage, so if the
program just does keep the stuff around, it will fill the memory, or the
disk, or anything, just keep that checkbox checked and walk away.
If you actually want help with leaks, better describe what you do and how
you got them. (Though you said you already got rid of them, it is hardly
your aim, probably you just test how much posts people here read before
plonk or something... )
How do you design such a thing?
ANY clue?
Gimmme your rough model and we'll see whose code is designed
"better".
And I NEVER EVER saw the issue of this kind with Java.
Better explain why java turns out to manage infinite amount of resources
better. ;-)
Is it like Chuck Norris, who did count to infinity? Twice! ?
I never had problems with memory leaks in C++.
I bow down to you, or immortal!
(***) > Except my lips are smiling.
What a fool they say in combination with my brain.
Guess if you saved up the trolling time to study well-known methods, you'd
be out of the swamp long ago. For sure, whining and blaming is easier.
(not like with C). Many
people here have the same experience. Yes, writing programs including
smilar to that you described. Just use a manager, like a smaprt pointer
consistently with every allocation. No leaks possible. (And if you
need
to keep blocks unreleased an indefinite time waiting user interaction --
that is not a leak issue, but consequence of a general design decision.)
Well, may be. If I ever find time to look at the design again,
who knows, I may even consider your "smart pointer" suggestion.
Just curious, you DID hear about existance of std::string, std::vector,
std::tr1:shared_ptr?
Dig?
next thing on my stack is
http://programmer.97things.oreilly.com/wiki/index.php/Edited_Contributions
and CERT's C++ section...
And to me, personally, the memory leak issues are some of the
top priority items and it is very unfortunate that this issue
is not resolved to this day in C++.
To my, and probably a big bunch of people this issue is considered as
soled
and non-issue for extremely long.
I am glad to hear that.
Unfortunately, I don't know you enough to see how much
what is you say is worth.
I'm sure. Your recent post suggest you hardly consirer anyone worth
listening to, including Sutter, Abrahams, so I'm not eager to fight for your
approval.
ALL I know is it is solved for what
I am doing in Java, which is my preferred language for several
years, and I am happy to hear that you think it is solved in C++
also, even though I have my doubts.
Java can collect only memory, with other resources you're almost as hosed as
in C. Well, you have finally to replace goto shutdown, still far from nice
and safe.
If it still hits you there must be some
other issue in the background. And beyond RAII, garbage collection is
also an option to be plugged in, if one really think it is needed or is
the
solution for a particular problem.
Cool. I did not know that. Not that I much care about it now.
Again, I am working in Java lately. Do not have much hope for
C++ and would not invest a penny into writing any new code in C++,
unless it is absolutely necessary, which I do not see so far.
So you admit that your presence here and the lengthy posts are just
trolling. ;)
Take for example the issue with writing a portable GUI code.
Is there such a thing?
Yep. Ther IS. Believe it or not.
It is called AWT or modern version is called swing.
Gives you 100% portable GUI code.
Sure, the fine details of the LOOk of some visual components
are going to be rendered slightly differently, especially in
light of Microsofts pattent issues. But the overall functionality
of your GUI is going to work 100% on any platform wher you can
install the JVM. Simple as that.
"generally portable" and "where JVM exists" is quite a diference.
The nice thing about Microsoft MVM is that is has about twice
better performance than Suns JVM, which is what I really
appreciate. Because when I do my processing, archive updates
and site generation, jobs run into hours, if not days.
It is a nightmare in C++ environment.
Quite so. And in any other if you actually mean portability.
Yes, I actually DO mean portability.
You just admitted you do not a few lines ago.
And it is a non issue with Java.
Sure. All you need is an already ported Java, and a big deal of resources.
I know java is preferred by many programmers -- I also know that the result
is so often hated by users. Even those having resources in abundance.
Sorry to tell you that.
You actually think I'm unaware what java can and can't do?
I saw a couple of projects that switched to java for that property (or
switched the interface, calling out for "work" written in whatever else),
experience shows the picture is far from that bright.
Well don't know where you get it from.
But it is bright enough for what I am interested in.
In fact, I'd say it is far superior to all the Microsoft ugly
hacks in GUI design, that cause nothing more than goose bumps
when I even think of it. I can implement a much better functionality
of GUI in Java in about 2 to 10 times less of a time,
doing ANYTHING I can imagine as far as GUI goes.
Just for a sanity check, you know what common cotrold are in Windows, and
what the listview control is?
And the "success" stories I saw have a really simple, and uncool
interface -- that looks pretty alien on at least one platform, if not on
all
of them.
Too bad. I don't happen to share your opinion.
If all you have is a hammer everything looks like a nail. ;-)
And I am not peddling Java here. Too bad everyone seems to be
abandoning it in a hurry. The traffic on Java collections of mine
are about 10 times as small as for MFC.
And, interestingly enough, C++ traffic is no more than
a half of MFC/VC/ATL/STL collection, which tells ME something.
Not sure if it tells YOU something though.
What traffic?
And I am not even using the swing version of GUI.
AWT is just fine for what I am doing and the user interface
is probably the best you can imagine in the wildest dreams.
I could care less about swing. Everything is just fine.
Great. So the next question is what on earth you do in this forum then?
Instead on living the happy life with java and AWT?
Do you mind?
:--}
Not really, just never understood the approach. I don't like java, but
express that by simply staying away. Not by going to java forums and start
flames. Either on made-up issues the real practicioners there know better
to be a non-issue, or real pains, that they also know ways better.
When I'm confident, I don;t need to prove it to anyone.
To tell you the truth, I do not know.
I just came here a couple of days ago for some small things.
One thing leads to another.
I saw some posts on programming issues.
So...
I think the issue of exceptions is significant enough,
and when someone asks for an opinion, do you mind if I have an
opinion?
Me? Absolutely not. ;) Opinions are like assholes, everyone has one, and
not very interested in others'...
Or you have nothing better to do in your life besides
poking your nose into affairs of others and questioning their
integrity?
see (***) ?
And it is a pitty that C++ can not satisfy either of these morst
critical elements for me.
Why should C++ do anything when you already found your silver environment?
Why turn C++ to clone an existing and established thing?
Ok, enough for this one.
Looks like a hopeless case indeed.
I don't deal with people, biased out of their head.
Cya.
LOL.