This is the mail archive of the libstdc++@gcc.gnu.org mailing list for the libstdc++ project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: Thanks to you for your work! (was: Re: Boost regression testsuite results for...)


Hi,

Gabriel Dos Reis wrote:

[snip, but good luck!]

> | - Unfortunately, whereas the libstdc++-v3 accompanying the current
> | Gcc3.0 works well for me, I cannot say so for that bundled with Gcc3.1
> | (PR2523). I'm really worried by this, because the newer hot stuff is in
> | the latter and my system is very common and stable :(
> |
> | - What about Loki ?!?
>
> Paolo, please could try a compilation and fill the output in GNATS?
> That would be a starting point.  I haven't yet had the time to look
> into Andrei's stuff and I'm not sure I'll have time to do it before
> next week or so.
>
> Thanks!

My pleasure!

In fact I already tried that (after having battled for a couple of days with
various typos and glitches in the distributed code, as per pme warning!)
And my *preliminary* results for main.cc (it includes and tests everything)
are not very good, indeed (session transcript follows):

--------
paolo:~/Programming/c++/Loki-mod> gcc -v
Reading specs from /usr/local/gcc/lib/gcc-lib/i686-pc-linux-gnu/3.0/specs
Configured with: ../gcc-20010416/configure --prefix=/usr/local/gcc
--disable-checking --enable-languages=c,c++,f77
gcc version 3.0 20010416 (prerelease)
paolo:~/Programming/c++/Loki-mod> g++ -o main main.cpp
In file included from Typelist.h:22,
                 from MultiMethods.h:21,
                 from main.cpp:3:
TypeManip.h:98: warning: all member functions in class `Loki::Conversion<T,
U>'
   are private
In file included from main.cpp:6:
SmartPtr.h:162: warning: `class Loki::RefCountedMT<P, ThreadingModel>' only
   defines private constructors and has no friends
SmartPtr.h:923: warning: all member functions in class `Loki::SmartPtr<T,
   OwnershipPolicy, ConversionPolicy, CheckingPolicy,
StoragePolicy>::Tester'
   are private
HierarchyGenerators.h: In instantiation of
`Loki::GenScatterHierarchy<DocElement,
Loki::Private::VisitorBinder<R>::Result>':
main.cpp:89:   instantiated from
`Loki::GenScatterHierarchy<Loki::Typelist<DocElement,
Loki::Typelist<Paragraph, Loki::Typelist<RasterBitmap, Loki::NullType> > >,
Loki::Private::VisitorBinder<R>::Result>'
main.cpp:89:   instantiated from `Loki::CyclicVisitor<void,
Loki::Typelist<DocElement, Loki::Typelist<Paragraph,
Loki::Typelist<RasterBitmap, Loki::NullType> > > >'
main.cpp:89:   instantiated from here
HierarchyGenerators.h:55: Internal error: Segmentation fault
Please submit a full bug report, with preprocessed source if appropriate.
See <URL:http://www.gnu.org/software/gcc/bugs.html> for instructions.
---------

However, please consider this as strictly *preliminary* (however, the seg
fault is *real* :(.
I'm planning to carefully "correct" the code again starting from the
distributed version and cross checking with you all any dubious
modification.
Only then file a GNATS PR, which, I hope, should also hint to the most
critical Loki libraries.

P.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]