[Bug libstdc++/51540] New: partial_sum (int *, int *, int *, multiplies < int >) does not use operator +(complex, complex)

[Bug libstdc++/51540] New: partial_sum (int *, int *, int *, multiplies < int >) does not use operator +(complex, complex)

[giecrilj at stegny dot 2a.pl]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51540

             Bug #: 51540
           Summary: partial_sum (int *, int *, int *, multiplies < int >)
                    does not use operator +(complex, complex)
    Classification: Unclassified
           Product: gcc
           Version: 4.5.1
               URL: http://gcc.gnu.org/onlinedocs/libstdc++/latest-doxygen
                    /a01567.html#a17bad49fa9fb1c5c41d019b16970fe9e
            Status: UNCONFIRMED
          Severity: major
          Priority: P3
         Component: libstdc++
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: giecrilj@stegny.2a.pl
              Host: i586-suse-linux
            Target: i586-suse-linux
             Build: i586-suse-linux


The following code fails to run:

int main () 
{ 
  int x [02] = { 0, 01 }; 
::std ::partial_sum (x, x + 02, x, ::std ::multiplies < int > ()); 
assert (+x [01] == 01); 
 return +EXIT_SUCCESS; }

Actual result: 
int main(): Assertion `+x [01] == 01' failed.

Expected result: 
I do not know, 
as the documentation does not say which complex class should be used.
Maybe it should not compile at all.

[Bug libstdc++/51540] partial_sum (int *, int *, int *, multiplies < int >) does not use operator +(complex, complex)

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51540

--- Comment #1 from Jonathan Wakely <redi at gcc dot gnu.org> 2011-12-13 23:41:26 UTC ---
what has complex got to do with it?

[Bug libstdc++/51540] partial_sum (int *, int *, int *, multiplies < int >) does not use operator +(complex, complex)

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51540

Jonathan Wakely <redi at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|                            |INVALID
           Severity|major                       |normal

--- Comment #2 from Jonathan Wakely <redi at gcc dot gnu.org> 2011-12-14 00:05:56 UTC ---
the behaviour's correct

[Bug libstdc++/51540] partial_sum (int *, int *, int *, multiplies < int >) does not use operator +(complex, complex)

[giecrilj at stegny dot 2a.pl]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51540

Christopher Yeleighton <giecrilj at stegny dot 2a.pl> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |UNCONFIRMED
         Resolution|INVALID                     |

--- Comment #3 from Christopher Yeleighton <giecrilj at stegny dot 2a.pl> 2011-12-14 06:22:19 UTC ---
(In reply to comment #1)
> what has complex got to do with it?

The documentation says this algorithm uses operator+; when you follow, it turns
out that it means operator + (complex, complex).

[Bug libstdc++/51540] partial_sum (int *, int *, int *, multiplies < int >) does not use operator +(complex, complex)

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51540

Paolo Carlini <paolo.carlini at oracle dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|                            |INVALID

--- Comment #4 from Paolo Carlini <paolo.carlini at oracle dot com> 2011-12-14 09:38:39 UTC ---
When you follow?!? Simply there are no complex numbers in your code. Full stop.

[Bug libstdc++/51540] partial_sum (int *, int *, int *, multiplies < int >) does not use operator +(complex, complex)

[giecrilj at stegny dot 2a.pl]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51540

Christopher Yeleighton <giecrilj at stegny dot 2a.pl> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |UNCONFIRMED
         Resolution|INVALID                     |

--- Comment #5 from Christopher Yeleighton <giecrilj at stegny dot 2a.pl> 2011-12-14 10:20:29 UTC ---
(In reply to comment #4)
> When you follow?!? Simply there are no complex numbers in your code. Full stop.

When you follow the hyperlink, you land at the description of operator +
(complex, complex).  When you hover, you are being told "returns new value
complex x + y" (not in Lynx though).

My code does not contain complex numbers but the algorithm is documented to use
an operator that uses complex numbers.

[Bug libstdc++/51540] partial_sum (int *, int *, int *, multiplies < int >) does not use operator +(complex, complex)

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51540

Paolo Carlini <paolo.carlini at oracle dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|                            |INVALID

--- Comment #6 from Paolo Carlini <paolo.carlini at oracle dot com> 2011-12-14 10:36:42 UTC ---
Hey, please do not re-open PRs at will, and let's not waste further time at
will. We know the documentation is **our** documentation after all. You are
mistaken, you are not using complex numbers  in your code, in particular you
are *not* providing *any* overloaded operator+ in your code. There is nothing
to fix here, if you are not convinced, just try any other implementation of the
Standard lib around.

[Bug libstdc++/51540] partial_sum (int *, int *, int *, multiplies < int >) does not use operator +(complex, complex)

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51540

--- Comment #7 from Jonathan Wakely <redi at gcc dot gnu.org> 2011-12-14 11:02:39 UTC ---
(In reply to comment #5)
> My code does not contain complex numbers but the algorithm is documented to use
> an operator that uses complex numbers.

That's just a doxygen bug, linking to the wrong operator+

If you want to report a documentation bug, say so, instead of being vague and
claiming the code is doing the wrong thing.

Your code fails to run because your assertion tests the wrong thing, not
because partial_sum is wrong.
If you don't know what partial_sum does, it's documented in many other places
e.g. http://www.sgi.com/tech/stl/partial_sum.html

[Bug libstdc++/51540] doxygen documentation for partial_sum misleading

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51540

Jonathan Wakely <redi at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |ASSIGNED
           Keywords|                            |documentation
   Last reconfirmed|                            |2011-12-14
         AssignedTo|unassigned at gcc dot       |redi at gcc dot gnu.org
                   |gnu.org                     |
         Resolution|INVALID                     |
     Ever Confirmed|0                           |1
            Summary|partial_sum (int *, int *,  |doxygen documentation for
                   |int *, multiplies < int >)  |partial_sum misleading
                   |does not use operator       |
                   |+(complex, complex)         |
           Severity|normal                      |minor

--- Comment #8 from Jonathan Wakely <redi at gcc dot gnu.org> 2011-12-14 11:05:04 UTC ---
I'll fix those docs, the overload taking a predicate doesn't even use operator+
anyway.

[Bug libstdc++/51540] doxygen documentation for partial_sum misleading

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51540

--- Comment #9 from Paolo Carlini <paolo.carlini at oracle dot com> 2011-12-14 11:17:44 UTC ---
Thanks for the docs, Jon.

[Bug libstdc++/51540] doxygen documentation for partial_sum misleading

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51540

Jonathan Wakely <redi at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.6.3

--- Comment #11 from Jonathan Wakely <redi at gcc dot gnu.org> 2011-12-15 09:51:12 UTC ---
Fixed for 4.7.0 so far

[Bug libstdc++/51540] doxygen documentation for partial_sum misleading

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51540

--- Comment #10 from Jonathan Wakely <redi at gcc dot gnu.org> 2011-12-15 09:50:04 UTC ---
Author: redi
Date: Thu Dec 15 09:49:59 2011
New Revision: 182359

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=182359
Log:
    PR libstdc++/51540
    * include/bits/stl_numeric.h (partial_sum): Adjust doxygen comments.

Modified:
    trunk/libstdc++-v3/ChangeLog
    trunk/libstdc++-v3/include/bits/stl_numeric.h

[Bug libstdc++/51540] doxygen documentation for partial_sum misleading

[giecrilj at stegny dot 2a.pl]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51540

--- Comment #12 from Christopher Yeleighton <giecrilj at stegny dot 2a.pl> 2011-12-16 18:59:31 UTC ---
Additionally, and also for the default operator form, it is unclear what the
result is when the operator is noncommutative.  That is, whether y[n+1] is set
to x[n+1]+y[n] or y[n]+x[n+1], or if the order is somewhat random altogether
(which would make partial_sum unusable as a potential driver for the equation
y[n+1]=f(y[n])).

[Bug libstdc++/51540] doxygen documentation for partial_sum misleading

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51540

--- Comment #13 from Paolo Carlini <paolo.carlini at oracle dot com> 2011-12-16 20:16:16 UTC ---
Up to Jon to decide which details he wants to add to the docs, but please
always keep in mind that in the areas where the behavior isn't implementation
defined only the ISO Standard (possibly amended per DRs) is the ultimate
reference.

[Bug libstdc++/51540] doxygen documentation for partial_sum misleading

[giecrilj at stegny dot 2a.pl]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51540

--- Comment #14 from Christopher Yeleighton <giecrilj at stegny dot 2a.pl> 2011-12-16 20:31:24 UTC ---
I would rather prefer to be able to use gcc (as a software developer) while not
having the ISO standard, which is 1) unreadable with an unarmed eye, 2) not
free (as in anything), 3) not bundled (of course), 4) written by and for
compiler manufacturers.  I understand there is no documentation for the
language itself, but that is (more or less) easily taught; however, requiring
the developer to know all the (essential) details of the standard library by
heart is a bit too much for me.  So I have taken the liberty of pointing out
this little inconvenience and I am trying to use the transform algorithm
instead.

[Bug libstdc++/51540] doxygen documentation for partial_sum misleading

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51540

--- Comment #15 from Jonathan Wakely <redi at gcc dot gnu.org> 2011-12-17 13:23:11 UTC ---
There are good books by Josuttis and Austern, or the SGI docs online. Our docs
could definitely improved, but there are other ways to get the info you want,
if you look for it

[Bug libstdc++/51540] doxygen documentation for partial_sum misleading

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51540

--- Comment #16 from Jonathan Wakely <redi at gcc dot gnu.org> 2011-12-17 13:26:04 UTC ---
Or of course you can look at the code

[Bug libstdc++/51540] doxygen documentation for partial_sum misleading

[giecrilj at stegny dot 2a.pl]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51540

--- Comment #17 from Christopher Yeleighton <giecrilj at stegny dot 2a.pl> 2011-12-17 17:05:58 UTC ---
The SGI documentation clarifies all issues well enough indeed.  I wonder
whether it would be a good thing to put a reference to SGI in the documentation
index (with a warning that it is somewhat out of date but still useful); the
SGI documentation is currently only accessible from the FAQ, and that is rather
hard to discover.

[Bug libstdc++/51540] doxygen documentation for partial_sum misleading

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51540

--- Comment #18 from Jonathan Wakely <redi at gcc dot gnu.org> 2011-12-18 22:33:19 UTC ---
Author: redi
Date: Sun Dec 18 22:33:15 2011
New Revision: 182461

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=182461
Log:
    PR libstdc++/51540
    * include/bits/stl_numeric.h (partial_sum): Adjust doxygen comments.

Modified:
    branches/gcc-4_6-branch/libstdc++-v3/ChangeLog
    branches/gcc-4_6-branch/libstdc++-v3/include/bits/stl_numeric.h

[Bug libstdc++/51540] doxygen documentation for partial_sum misleading

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51540

Jonathan Wakely <redi at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED

--- Comment #19 from Jonathan Wakely <redi at gcc dot gnu.org> 2011-12-18 22:34:32 UTC ---
also fixed in the 4.6 branch