::std sometimes needed, sometimes not

[Marcin Vorbrodt]

Hi there. How come when in one file i unclude <cmath> i need to use
std::tan(...), and in another file i include <cassert> and std::assert fails
to compile. Nowhere in my code i do using namespace std; Seams like some
things in my header files are in std namespace and some are not. Is that
correct, or is my compiler (mingw, 3.2 i think) horribly broken? )

Thanx,
Martin

 

[Alf P. Steinbach]

Hi there. How come when in one file i unclude <cmath> i need to use
std::tan(...), and in another file i include <cassert> and std::assert fails
to compile.

That's because assert is a macro. It doesn't live in a namespace; no
macro does. That's one problem with macros; the nice thing about them
is that they can pick up things like line number and so on.

Nowhere in my code i do using namespace std; Seams like some
things in my header files are in std namespace and some are not. Is that
correct

Yes, but there are only a handful or so of standard macros.

or is my compiler (mingw, 3.2 i think) horribly broken? )

Wrt. to some other things, yes, but it's still a very good compiler.

 

[Marcin Vorbrodt]

OK, so is there a "C++" way of doing assert? A standard C++ class/template
or something that does assert? Also, is there a C++ alternative to <cmath>?

Thanx,
Martin

 

[Kevin Goodsell]

OK, so is there a "C++" way of doing assert? A standard C++ class/template
or something that does assert? Also, is there a C++ alternative to <cmath>?

Please don't top-post here. Read section 5 of the FAQ for posting
guidelines:

http://www.parashift.com/c++-faq-lite/

C++ doesn't have anything assert-like built in (other than assert), but
you could certainly create something, probably using an exception class
to carry the information about the error. Stroustrup has an example of
this somewhere, IIRC. Probably in "The C++ Programming Language".

<cmath> *is* the C++ alternative.

-Kevin

 

[llewelly]

Please don't top-post here. Read section 5 of the FAQ for posting
guidelines:

http://www.parashift.com/c++-faq-lite/

C++ doesn't have anything assert-like built in (other than assert),
but you could certainly create something, probably using an exception
class to carry the information about the error.

Before you write code signaling a programmer error with an exception,
read this thread on assert versus exceptions:

http://groups.google.com/[email protected]
or:
http://tinyurl.com/nhp7

Stroustrup has an
example of this somewhere, IIRC. Probably in "The C++ Programming
Language".

[snip]

He does. However, other experts have since argued that is generally
bad advice.

 

[Kevin Goodsell]

Before you write code signaling a programmer error with an exception,
read this thread on assert versus exceptions:

http://groups.google.com/[email protected]
or:
http://tinyurl.com/nhp7

I don't know if you were talking specifically about the particular
message that link goes to, but I don't think I could disagree more with
this point:

- You definitely do NOT want to throw an exception in case of a
contract violation. It is very, very rare, if ever, that this is
correct. The program is incorrect, and you abort.

The fact that every function can potentially throw an exception
makes writing exception safe code pratically impossible.

A lot of people seem to want exceptions anyway. To keep them
happy, I would suggest some sort of callback, which the user can
replace. By default, the code should abort. If the user replaces
it with something which will throw, his code probably will not work
when the exception occurs. But since that's what he wanted...

This sounds like someone who doesn't understand exceptions to me.

Stroustrup has an
example of this somewhere, IIRC. Probably in "The C++ Programming
Language".

[snip]

He does. However, other experts have since argued that is generally
bad advice.

Reference? Without reading the opposing argument, I'm very likely to
side with Stroustrup over some generic, unnamed "experts".

-Kevin

 

[Gianni Mariani]

llewelly wrote:
...

- You definitely do NOT want to throw an exception in case of a
contract violation. It is very, very rare, if ever, that this is
correct. The program is incorrect, and you abort.

The fact that every function can potentially throw an exception
makes writing exception safe code pratically impossible.

A lot of people seem to want exceptions anyway. To keep them
happy, I would suggest some sort of callback, which the user can
replace. By default, the code should abort. If the user replaces
it with something which will throw, his code probably will not work
when the exception occurs. But since that's what he wanted...

This sounds like someone who doesn't understand exceptions to me.

OK - time for some friendly debate.

I'll state that I use exceptions very rarely and hardly ever in
production code.

I actually agree the the quote 100% because I have done exactly that and
found that it is a very acceptable paradigm for developing solid and
easily maintainable code.

I have found that exceptions leed to sloppy code by less experienced
programmers and even some more experienced guys as well.

However, I'm open to suggestions. I'm only 90% conviced they're next to
useless. OK OK OK. In some rare circumstances I can see that they are
an acceptable alternative. In a test framework, I designed, throwing an
exception was a way to indicate a test failure, abort()ing was
unsatisfactory since it meant that other tests would have not been able
to run in that particular framework.

 

[Jonathan Mcdougall]

- You definitely do NOT want to throw an exception in case of a
contract violation. It is very, very rare, if ever, that this is
correct. The program is incorrect, and you abort.

The fact that every function can potentially throw an exception
makes writing exception safe code pratically impossible.

A lot of people seem to want exceptions anyway. To keep them
happy, I would suggest some sort of callback, which the user can
replace. By default, the code should abort. If the user replaces
it with something which will throw, his code probably will not work
when the exception occurs. But since that's what he wanted...

This sounds like someone who doesn't understand exceptions to me.

However, I'm open to suggestions. I'm only 90% conviced they're next to
useless. OK OK OK. In some rare circumstances I can see that they are
an acceptable alternative. In a test framework, I designed, throwing an
exception was a way to indicate a test failure, abort()ing was
unsatisfactory since it meant that other tests would have not been able
to run in that particular framework.

I use exceptions primarily for signaling constructor failure, since it
is the only way. Do you have a workaround or you assume your ctors
always succeed ?

Jonathan

 

[Gianni Mariani]

- You definitely do NOT want to throw an exception in case of a
contract violation. It is very, very rare, if ever, that this is
correct. The program is incorrect, and you abort.

The fact that every function can potentially throw an exception
makes writing exception safe code pratically impossible.

A lot of people seem to want exceptions anyway. To keep them
happy, I would suggest some sort of callback, which the user can
replace. By default, the code should abort. If the user replaces
it with something which will throw, his code probably will not work
when the exception occurs. But since that's what he wanted...

This sounds like someone who doesn't understand exceptions to me.

However, I'm open to suggestions. I'm only 90% conviced they're next to
useless. OK OK OK. In some rare circumstances I can see that they are
an acceptable alternative. In a test framework, I designed, throwing an
exception was a way to indicate a test failure, abort()ing was
unsatisfactory since it meant that other tests would have not been able
to run in that particular framework.

I use exceptions primarily for signaling constructor failure, since it
is the only way. Do you have a workaround or you assume your ctors
always succeed ?

Very few contructors I create fail. bad_alloc is a given but I usually
have bigger problems than bad_alloc exceptions. (I think I have yet to
see a bad_alloc exception thrown in a production environment) With most
systems now allowing you to overcommit on memory, you're more likely to
get an access violation/segmentation fault when memory is depleted.

I'll give you that one, but I think it's rare and hence letting it abort
may be a suitable alternative most of the time.

Anything else ?

 

[Jonathan Mcdougall]

However, I'm open to suggestions. I'm only 90% conviced they're next to

Very few contructors I create fail.

That's what I thought

bad_alloc is a given but I usually
have bigger problems than bad_alloc exceptions.

And you usually abort on these problems?

(I think I have yet to
see a bad_alloc exception thrown in a production environment)

Happens often when you try to run recent software on old machines,
believe me

With most
systems now allowing you to overcommit on memory, you're more likely to
get an access violation/segmentation fault when memory is depleted.

Mmm.. I didn't know that.

I'll give you that one, but I think it's rare and hence letting it abort
may be a suitable alternative most of the time.

I think aborting is never a suitable alternative. Letting a catchable
exception propagate is imho bad programming. Trying something else
(or the same thing), warning the user, asking to save.. you know..
user-friendly programming. Nothing frustrates me more than a program
terminating without warning.

I long believed my programs would never fail as others do, but now I
try to do "preventive programming".

I understand there are problems you just cannot foresee or even solve
during runtime, but I try to do as much as possible to prevent my
applications from crashing.


Jonathan

 

[Gianni Mariani]

Jonathan Mcdougall wrote:
....

And you usually abort on these problems?
Yep.



Happens often when you try to run recent software on old machines,
believe me

Hmm. I have developed software that ran on thousands of machines and
other problems were reported, but out of memory errors were not among them.

Mmm.. I didn't know that.

I may have overstated that a bit, but I have seen this.

I think aborting is never a suitable alternative. Letting a catchable
exception propagate is imho bad programming.

On some systems (Unix) you get a nice big fat core file pointing you to
the code and the reason for the fault. Sounds like a wonderful
alternative to me !

.... Trying something else

(or the same thing), warning the user, asking to save.. you know..
user-friendly programming. Nothing frustrates me more than a program
terminating without warning.

Sometimes there are more important features. If you can solve the
problem with a memory upgrade, it's often cheaper than complicating your
code.

I long believed my programs would never fail as others do, but now I
try to do "preventive programming".

I understand there are problems you just cannot foresee or even solve
during runtime, but I try to do as much as possible to prevent my
applications from crashing.

Not letting them happen in the first place is the key.

There are ways to guarentee that you can't get a bad_alloc but it
requires knowing before-hand what you're going to need - essentially
pre-allocating the memory.

Seems like alot of trouble for not alot of gain.

If I'm not mistaken, it sounds like we may have a different bias but we
agree on the problem.

 

[Kevin Goodsell]

- You definitely do NOT want to throw an exception in case of a
contract violation. It is very, very rare, if ever, that this is
correct. The program is incorrect, and you abort.

The fact that every function can potentially throw an exception
makes writing exception safe code pratically impossible.

A lot of people seem to want exceptions anyway. To keep them
happy, I would suggest some sort of callback, which the user can
replace. By default, the code should abort. If the user replaces
it with something which will throw, his code probably will not work
when the exception occurs. But since that's what he wanted...

This sounds like someone who doesn't understand exceptions to me.

OK - time for some friendly debate.

I'll state that I use exceptions very rarely and hardly ever in
production code.

I actually agree the the quote 100% because I have done exactly that and
found that it is a very acceptable paradigm for developing solid and
easily maintainable code.

I don't think that library code should ever unilaterally abort the
program. Besides that, doing so causes objects to go undestructed,
leading to resource leaks (sure, most modern desktop systems clean up
when the process terminates, but you can't count on that in portable code).

I don't understand why you would prefer aborting to throwing an
exception anyway. What's the worst case for throwing an exception? It's
not caught, and the program aborts. So by aborting instead of throwing,
you're ensuring that you always get the worst case.

I have found that exceptions leed to sloppy code by less experienced
programmers and even some more experienced guys as well.

However, I'm open to suggestions. I'm only 90% conviced they're next to
useless. OK OK OK. In some rare circumstances I can see that they are
an acceptable alternative. In a test framework, I designed, throwing an
exception was a way to indicate a test failure, abort()ing was
unsatisfactory since it meant that other tests would have not been able
to run in that particular framework.

Well, I have to admit that I don't have a lot of experience actually
using exception handling, but I like it a lot. I find that it makes code
considerably easier to write, and less cluttered. During development of
a library, I just catch everything in main and print the what() string.
That way when I run a quick test I get a nice little report of what went
wrong (if anything). In a program using the library I just catch
wherever is most appropriate.

I don't know what else to even say about it. It makes things so simple I
can't imagine why you wouldn't want to use it. I hate having to
constantly check return values for errors. Speaking of sloppy
programming, how often do people forget to check return values? It's a
bit harder to ignore an exception, and easier to track down the
resulting problem if you do. Ignore a return value and your program may
limp along in an undefined state for a while before crashing in some
other part of the program. With an uncaught exception it's immediate,
and the debugger can show you where it was thrown from.

I might have to ask you to explain to me what's so bad about exceptions. :-/

-Kevin

 

[Jerry Coffin]

[ commenting on a quote from James Kanze ... ]

This sounds like someone who doesn't understand exceptions to me.

Others would disagree. Just for example, Herb Sutter (author of
Exceptional C++, More Exceptional C++ and an upcoming and TTBOMK, as yet
unnamed book on exceptions) lists James Kanze as "guru" here:

http://www.gotw.ca/gotw/056.htm

[ ... ]

Reference? Without reading the opposing argument, I'm very likely to
side with Stroustrup over some generic, unnamed "experts".

I'm not sure of the arguments in question either, but if James Kanze is
among those making the argument, I'd guess that at the very least, it's
worth studying in detail. Herb Sutter isn't the only one who names or
acknowledges him either -- he is, for example, listed in the
acknowledgements of a number of books I have handy, including (for one
example) _C++ In a Nutshell_, from O'Reilly and Associates (who
generally aren't disposed toward publishing really bad books anyway).

 

[Jerry Coffin]

[ ... ]

I don't understand why you would prefer aborting to throwing an
exception anyway. What's the worst case for throwing an exception? It's
not caught, and the program aborts. So by aborting instead of throwing,
you're ensuring that you always get the worst case.

A contrast violation means there is something wrong with the code (i.e.
a bug) that has been found and noted, and needs to be fixed.

In this case, aborting is a LONG ways from being the worst case. Quite
the contrary: the worst case is that the problem is allowed to continue
unnoticed for some arbitrary period of time -- almost inevitably, the
longer a bug persists, the more difficult and expensive it is to fix.
Therefore, when you DO find what you're sure is a bug in some code, the
best case is to ensure that it fails as quickly, dependably and
spectacularly as possible, so it will be fixed as soon as possible
instead of being allowed to persist.

The problem with throwing an exception in a case like this is that it's
all too easy (and common) for somebody to do something like this:

int main(int argc, char **argv) {
do {
try {
return real_main(argc, char **argv);
}
catch(...) {
std::cerr<<"Some problem -- retrying\n";
}
} while(1);
return 0; // never really executes
}

In fact, almost any instance of "catch(...)" can cover up an almost
arbitrary collection of problems that the library designer is TRYING to
make sure get noticed.

One way to avoid that is to abort instead of throwing an exception -- as
long as you're sure that what you've got is a bug in the code, and not
something from which recovery is reasonable, abort() is a better option
precisely because it prevents any other part of the code from covering
up the problem.

 

[Attila Feher]

[ commenting on a quote from James Kanze ... ]

This sounds like someone who doesn't understand exceptions to me.

Others would disagree. Just for example, Herb Sutter (author of
Exceptional C++, More Exceptional C++ and an upcoming and TTBOMK, as
yet unnamed book on exceptions) lists James Kanze as "guru" here:

James is a guru. A practical one. So some of his advice is not about how
to do things in a perfect world where all compilers are conforming to the
standard, but how to do things in the real world. He is a top consultant
and rather a pragmatic, real-world guy than a language lawyer. It is worth
to listen to him.

 

[Buster Copley]

[ commenting on a quote from James Kanze ... ]

This sounds like someone who doesn't understand exceptions to me.

Others would disagree. Just for example, Herb Sutter (author of
Exceptional C++, More Exceptional C++ and an upcoming and TTBOMK, as yet
unnamed book on exceptions) lists James Kanze as "guru" here:

http://www.gotw.ca/gotw/056.htm

[ ... ]

Reference? Without reading the opposing argument, I'm very likely to
side with Stroustrup over some generic, unnamed "experts".

I'm not sure of the arguments in question either, but if James Kanze is
among those making the argument, I'd guess that at the very least, it's
worth studying in detail. Herb Sutter isn't the only one who names or
acknowledges him either -- he is, for example, listed in the
acknowledgements of a number of books I have handy, including (for one
example) _C++ In a Nutshell_, from O'Reilly and Associates (who
generally aren't disposed toward publishing really bad books anyway).

Heh heh. So am I, but I hope no one thinks I'm an expert.
Regards,
Buster.

 

[Kevin Goodsell]

A contrast violation means there is something wrong with the code (i.e.
a bug) that has been found and noted, and needs to be fixed.

For the record, I wasn't really considering the context of the quoted
message, just exception handling in general.

Now, if you were to use exception handling, and base your exceptions off
the standard exceptions, a contract violation would be considered a
std::logic_error. There are other types of exceptions that you might use
that would not be the result of a contract violation.

But looking at just the contract violation case, obviously it's a bug
that should be fixed. But at least some bugs always make it into the
final product. Terminating the program when a problem is found is not a
very useful behavior from the perspective of the user of the
application. In fact, I don't think most users would be very happy with
that.

In this case, aborting is a LONG ways from being the worst case. Quite
the contrary: the worst case is that the problem is allowed to continue
unnoticed for some arbitrary period of time

Yes, I agree that this is bad. But it only happens if you explicitly
ignore the exception.

-- almost inevitably, the
longer a bug persists, the more difficult and expensive it is to fix.
Therefore, when you DO find what you're sure is a bug in some code, the
best case is to ensure that it fails as quickly, dependably and
spectacularly as possible, so it will be fixed as soon as possible
instead of being allowed to persist.

This is not true from the user's perspective, though.

The problem with throwing an exception in a case like this is that it's
all too easy (and common) for somebody to do something like this:

int main(int argc, char **argv) {
do {
try {
return real_main(argc, char **argv);
}
catch(...) {
std::cerr<<"Some problem -- retrying\n";
}
} while(1);
return 0; // never really executes
}

In fact, almost any instance of "catch(...)" can cover up an almost
arbitrary collection of problems that the library designer is TRYING to
make sure get noticed.

Yes, I definitely agree that this is a problem. But one way to look at
it is that it's not really *your* problem. If you threw the exception,
you've done your part. If the caller ignores it, then they screwed up.
But this doesn't make any difference in many cases. If the program
breaks for the customer, you'll still be in trouble.

So I see the point, but I don't think aborting is an ideal solution, in
particular once the program is in the customer's hands. An exception
could be an ideal solution, if it's not ignored. As far as I know,
there's no way to be sure of this.

A few ideas come to mind, but none is anywhere near perfect. You could
abort or throw depending on a build setting, for example. But if you
test with abort and ship with throw then you don't even know if the
catching code is in place or doing anything sensible. You'd have to test
with throw, and it could be covered up.

You could try to make an un-ignorable exception, maybe. Suppose your
exception object includes a 'handled' flag. On construction, this flag
is false. The catcher must set it to true. On destruction, if the flag
does not indicate that the exception was handled, the destructor aborts
the program. The problem is that it can still be ignored, it just takes
a little more effort (catch, set the flag, do nothing else).

-Kevin

 

[Mike Wahler]

Hi there. How come when in one file i unclude <cmath> i need to use
std::tan(...), and in another file i include <cassert> and std::assert fails
to compile. Nowhere in my code i do using namespace std; Seams like some
things in my header files are in std namespace and some are not.

Exactly. Macros are not ('assert' is a macro) in a namespace.
They cannot be, since they're merely text substitutions.

Is that
correct, or is my compiler (mingw, 3.2 i think) horribly broken? )

It's fine.

-MIke

 

[llewelly]

I don't know if you were talking specifically about the particular
message that link goes to, but I don't think I could disagree more
with this point:

That's exactly the point I was referring to.

- You definitely do NOT want to throw an exception in case of a
contract violation. It is very, very rare, if ever, that this is
correct. The program is incorrect, and you abort.

The fact that every function can potentially throw an exception
makes writing exception safe code pratically impossible.

A lot of people seem to want exceptions anyway. To keep them
happy, I would suggest some sort of callback, which the user can
replace. By default, the code should abort. If the user replaces
it with something which will throw, his code probably will not work
when the exception occurs. But since that's what he wanted...

This sounds like someone who doesn't understand exceptions to me.

Stroustrup has an
example of this somewhere, IIRC. Probably in "The C++ Programming
Language".

[snip]
He does. However, other experts have since argued that is generally
bad advice.

Reference? Without reading the opposing argument, I'm very likely to
side with Stroustrup over some generic, unnamed "experts".

Here's David Abrahams, with a similar opinion:
http://www.boost.org/more/error_handling.html

As for 'unnamed experts', I felt it was obvious that the name of the
expert I was referring to was James Kanze. I won't claim his
opinions should always take precedence over Stroustrup's, but
IMO, James's posts show plenty of wisdom, experience, and
insight. (With some errors here and there. You should google
for his posts, and read a few to form your own opinion.

Finally, you could email Stroustrup for his current opinion on
whether exceptions should be used to signal programmer
errors. Maybe it has changed.

 

[llewelly]

llewelly wrote:
..

- You definitely do NOT want to throw an exception in case of a
contract violation. It is very, very rare, if ever, that this is
correct. The program is incorrect, and you abort.
The fact that every function can potentially throw an exception
makes writing exception safe code pratically impossible.
A lot of people seem to want exceptions anyway. To keep
them
happy, I would suggest some sort of callback, which the user can
replace. By default, the code should abort. If the user replaces
it with something which will throw, his code probably will not work
when the exception occurs. But since that's what he wanted...

This sounds like someone who doesn't understand exceptions to me.

OK - time for some friendly debate.

I'll state that I use exceptions very rarely and hardly ever in
production code.

I actually agree the the quote 100% because I have done exactly that
and found that it is a very acceptable paradigm for developing solid
and easily maintainable code.

I have found that exceptions leed to sloppy code by less experienced
programmers and even some more experienced guys as well.

However, I'm open to suggestions. I'm only 90% conviced they're next
to useless.

[snip]

Hm. Maybe it isn't clear from this snippet, but from what I know,
James isn't conviced exceptions are useless. He just thinks they
should be used for conditions one plans on responding to, and
*not* for programmer errors. If you read further into the
referenced thread, you'll see more.