Not GTG (was Re: PING: Proposal for Boost 1.33.0 package)

[Corinna Vinschen]

On Aug 20 13:02, Corinna Vinschen wrote:
> On Aug 20 12:39, Vaclav Haisman wrote:
> > On Fri, 19 Aug 2005, Charles Wilson wrote:
> > >Corinna Vinschen wrote:
> > >>I just had a look into the boost packages and from my point of view they
> > >>are not GTG, for various reasons.
> > >>- The Cygwin naming convention of DLLs is not used:
> > Boost obviously doesn't generate cyg- prefixed names, that would have to be 
> > fixed. But I can see precedence for not doing so in my /bin too. There are 
> > libpython2.4.dll, libW11.dll and libzsh-4.2.4.dll.
> 
> I don't think that the few exceptions from the rule should be used
> as excuse.  Rather they should also change their name in later versions.
> 
> > >>- The naming convention for static and dynamic link libs is not used:
> > The naming convention is different because Boost generates or can generate 
> > many combinations of libs. Static/dynamic, multi threaded/single threaded, 
> > debug/release, with or witout debug info. I do not think it would be a good 
> > idea to try to force different naming convention just for Cygwin.
> 
> Sorry, but that's not what I mean.  Regardless of the internal specification,
> the link libs should use the naming convention for static and dynamic libs,
> as it is automagically created by libtool.  I don't know about the internals
> of boost, so I assumed that the -s.a vs. .a was supposed to mean static vs.
> dynamic.  If it means something else, then the naming convention is still to
> use .a and .dll.a as in:
> 
>   libboost-foo-bar.a
>   libboost-foo-bar-s.a
>   libboost-foo-bar.dll.a
>   libboost-foo-bar-s.dll.a

Oh, sorry, I just reread your mail.  I missed that boost didn't build
the import libs at all.  I missunderstood this one.


Corinna

-- 
Corinna Vinschen                  Please, send mails regarding Cygwin to
Cygwin Project Co-Leader          mailto:cygwin@cygwin.com
Red Hat, Inc.