Apache and standard C++ library

Jonathan Wakely cow@compsoc.man.ac.uk
Tue May 17 10:02:00 GMT 2005


On Mon, May 16, 2005 at 05:19:12PM -0700, Joe Buck wrote:

> On Mon, May 16, 2005 at 02:40:16PM -0400, Andrew Pinski wrote:
> > I thought I would forward this along just in case nobody has seen it  
> > yet:
> > http://mail-archives.apache.org/mod_mbox/incubator-general/ 
> > 200505.mbox/ 
> > %3cC398496DF88AD711908C00065BEEE32C01473747@dfwmail.quovadx.com%3e
> 
> I could not follow that link, try
> 
> http://mail-archives.apache.org/mod_mbox/incubator-general/200505.mbox/%3cC398496DF88AD711908C00065BEEE32C01473747@dfwmail.quovadx.com%3e

Thanks for the links.

> > PS, I think libstdc++v3 should really become Apache's standard C++  
> > library
> > instead of RogueWave's "new" open source implementation.
> 
> I think Apache should not get into the business of supplying a standard
> C++ library; this appears to be an attempt by Rogue Wave to increase their
> business or help with their maintainance burden, or both.  If this

My thoughts exactly.

If Apache really need a non-GPL, portable stdlib what's wrong with STLPort ?

> proposal moves forward, it has the potential to break C++ on GNU/Linux
> (two binary-incompatible flavors of C++ library in wide use).
> 
> Apache should write C++ programs in C++ and report bugs to the compiler
> suppliers.  They should not write them to any particular C++ standard
> library implementation.

The ASF seem to have a critical case of NIH syndrome, if the code isn't
in their repository they won't use it.

> On the other hand, if it results in more free software, it's not such a
> bad thing.

True, but the announcement talks about "the a [sic] true standard
implementation that is also free and commonly available" as though
libstdc++ and STLPort don't exist or can't be considered standard.
Mudslinging doesn't help the free software community.

> But this is arguably a political problem, not an R&D problem; not really
> something we can do anything about on this list.

We can keep working to make libstdc++ as good as possible, but that
might not help if the ASF and RogueWave use their marketing budgets to 
make us look irrelevant.  We should at least be aware of the fact and
try to ensure libstdc++ will still be recognised.

Still, maybe the project will turn out to be as useful as Xerces ;)

jon



More information about the Libstdc++ mailing list