stdcxx-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Martin Sebor <se...@roguewave.com>
Subject Re: Problem building stdcxx on Windows XP
Date Fri, 01 Sep 2006 16:15:35 GMT
Geoffrey Winn wrote:
> On 31/08/06, Martin Sebor <sebor@roguewave.com> wrote:
> 
>>
>> Geoffrey Winn wrote:
>>
>> >
>> > 2. Once everything had completed, I can see in the Task List that I 
>> have
>> 8
>> > warnings and 18 errors.
>> >
>> > Are these results expected? If not, would you like the details?
>>
>> We are most likely aware of the issues but I would be happy to
>> go over the results and explain the failures if that would
>> increase your confidence in the project :) It may not be a bad
>> idea for us to document them in the form of Jira issues.
>>
>>
> Don't worry about explaining the issues - I trust you :-) My main concern
> was that the errors might indicate something I had done wrong.

I don't think so, they all look legit.

> The task 
> list
> shows the following. Apologies for the formatting, if it isn't readable let
> me know and I'll try another approach.
> 
> C:\Users\gwinn\Tuscany\stdcxx\stdcxx-4.1.3\examples\tutorial\dynatype.cpp(109): 
> 
> error C2248: 'dynatype::map<T>' : cannot access private struct declared in
> class 'dynatype'
>        with
>        [
>            T=void
>        ]
> C:\Users\gwinn\Tuscany\stdcxx\stdcxx-4.1.3\examples\tutorial\dynatype.cpp(114): 
> 
> error C2248: 'dynatype::map<T>' : cannot access private struct declared in
> class 'dynatype'
>        with
>        [
>            T=void
>        ]
> C:\Users\gwinn\Tuscany\stdcxx\stdcxx-4.1.3\examples\tutorial\dynatype.cpp(188): 

As Liviu already mentioned, compilers tend to have problems with
this example. It's an MSVC bug. I filed it in Jira and will report
it Microsoft as soon as their web site for reporting bugs works
again: http://issues.apache.org/jira/browse/STDCXX-282.

[...]
> C:\Users\gwinn\Tuscany\stdcxx\stdcxx-4.1.3\src\wcodecvt.cpp(385): warning
> C4307: '*' : integral constant overflow

This is also caused by a compiler bug. Here's a small test case:
   http://issues.apache.org/jira/browse/STDCXX-283


> C:\Users\gwinn\Tuscany\stdcxx\stdcxx-
> 4.1.3\tests\localization\22.locale.num.get.cpp(1582): warning C4005: 'LL' :
> macro redefinition

This is a bug in the test. It's fixed on trunk:
   http://svn.apache.org/viewvc?view=rev&revision=386904

[...]
> C:\Users\gwinn\Tuscany\stdcxx\stdcxx-4.1.3\tests\numerics\26.c.math.cpp(208): 
> 
> warning C4244: 'initializing' : conversion from 'double' to 'const float',
> possible loss of data

This is benign. We should probably silence it.

[...]
> C:\Users\gwinn\Tuscany\stdcxx\stdcxx-4.1.3\tests\src\printf.cpp(2275):
> warning C4244: '=' : conversion from 'unsigned __int64' to 'char', possible
> loss of data

Also benign but possibly non-trivial to silence since it's in
template code.

> C:\Users\gwinn\Tuscany\stdcxx\stdcxx-4.1.3\tests\src\printf.cpp(2275):
> warning C4244: '=' : conversion from 'wchar_t' to 'char', possible loss of
> data
> C:\Users\gwinn\Tuscany\stdcxx\stdcxx-4.1.3\tests\src\valcmp.cpp(798):
> warning C4244: 'return' : conversion from 'const IntT' to 'int', possible
> loss of data

Same here although I think it might have been fixed on trunk with
this change: http://svn.apache.org/viewvc?view=rev&revision=366950.

Martin

Mime
View raw message