Windows and apache stdcxx
Hi! I still have no reply on my mail from Fr 15.09.2006 11:24 regarding "Windows Boost with apache stdcxx". I'm really kind of stuck with this. Please anybody give me a hint on what to do with this. Thanks, Markus
Duft Markus wrote:
I still have no reply on my mail from Fr 15.09.2006 11:24 regarding "Windows Boost with apache stdcxx". I'm really kind of stuck with this. Please anybody give me a hint on what to do with this.
Since this seems to mainly be a build problem, the boost build list might have been a better place to ask. Details at: http://lists.boost.org/mailman/listinfo.cgi/boost-build Although, as far as I know, no one is currently using stdcxx with boost. So there isn't really any support available. Sorry. And if you do get the build sorted out, there's no guarantee that the libraries will work with stdcxx. Perhaps, STLport would be a better option? Anyway, your question:
I'm using boost for quite a while now, and ran into a problem. I'm writing a compiler wrapper for cl (wgcc, @ sourceforge) which behaves like gcc (wgcc runs under Interix). now i'm building tons of packages with wgcc. Due to problems with the microsoft STL we switched to apaches STDCXX implementation (wgcc uses it automatically for every C++ package built with it, if it is present). The problem now is, that all packages are built with stdcxx, EXCEPT of course boost, because i'm using the windows build mechanism, hence cl directly and not my wrapper, to build it. I haven't tried out yet, but i really think that this is a bad combination somehow. Have you got any advice on how to get boost built with the stdcxx library??
I can think of two possible solutions. You could use the gcc toolset with your wrapper. Perhaps by writing in your 'user-config.jam': using gcc : wgcc : c:/path/wgcc ; Putting the location of the wgcc executable as the last argument. Or maybe be writing a stdcxx toolset (which could be based on the stlport toolset, at tools/build/v2/tools/stlport.jam). Daniel
Daniel James wrote:
Duft Markus wrote:
I still have no reply on my mail from Fr 15.09.2006 11:24 regarding "Windows Boost with apache stdcxx". I'm really kind of stuck with this. Please anybody give me a hint on what to do with this.
Since this seems to mainly be a build problem, the boost build list might have been a better place to ask. Details at:
http://lists.boost.org/mailman/listinfo.cgi/boost-build
Although, as far as I know, no one is currently using stdcxx with boost. So there isn't really any support available. Sorry. And if you do get the build sorted out, there's no guarantee that the libraries will work with stdcxx. Perhaps, STLport would be a better option?
As I understand it, Apache stdcxx is derived from the Rogue Wave standard library. This library was used by the Solaris and perhaps some other compiler vendors. There are a set of configuration macros supplied in boost.config supplied in the roguewave.hpp setup. Various libraries have workarounds that are specific to this library. So the first step would probably be to use these for your initial configuration setup. Perhaps boost.config should be changed to recognize the apache lib so it can be supported more directly in the future. HTH, Jeff
Jeff Garland wrote:
As I understand it, Apache stdcxx is derived from the Rogue Wave standard library. This library was used by the Solaris and perhaps some other compiler vendors.
HP is one those vendors: HP-UX, Tru64 and VMS all use RW library. Boris ----- Original Message ----- From: "Jeff Garland" <jeff@crystalclearsoftware.com> To: <boost@lists.boost.org> Sent: Saturday, September 23, 2006 3:22 PM Subject: Re: [boost] Windows and apache stdcxx
Daniel James wrote:
Duft Markus wrote:
I still have no reply on my mail from Fr 15.09.2006 11:24 regarding "Windows Boost with apache stdcxx". I'm really kind of stuck with this. Please anybody give me a hint on what to do with this.
Since this seems to mainly be a build problem, the boost build list might have been a better place to ask. Details at:
http://lists.boost.org/mailman/listinfo.cgi/boost-build
Although, as far as I know, no one is currently using stdcxx with boost. So there isn't really any support available. Sorry. And if you do get the build sorted out, there's no guarantee that the libraries will work with stdcxx. Perhaps, STLport would be a better option?
As I understand it, Apache stdcxx is derived from the Rogue Wave standard library. This library was used by the Solaris and perhaps some other compiler vendors. There are a set of configuration macros supplied in boost.config supplied in the roguewave.hpp setup. Various libraries have workarounds that are specific to this library. So the first step would probably be to use these for your initial configuration setup. Perhaps boost.config should be changed to recognize the apache lib so it can be supported more directly in the future.
HTH,
Jeff _______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
Boris Gubenko wrote:
Jeff Garland wrote:
As I understand it, Apache stdcxx is derived from the Rogue Wave standard library. This library was used by the Solaris and perhaps some other compiler vendors.
HP is one those vendors: HP-UX, Tru64 and VMS all use RW library.
Right, the RW lib should be reasonably well supported by Boost.Config "as is", I think the issue was how to build with the Apache lib when it's *not* the one that comes with your compiler, which I suspect is a bit trickier. John.
John Maddock wrote:
Boris Gubenko wrote:
Jeff Garland wrote:
As I understand it, Apache stdcxx is derived from the Rogue Wave standard library. This library was used by the Solaris and perhaps some other compiler vendors. HP is one those vendors: HP-UX, Tru64 and VMS all use RW library.
Right, the RW lib should be reasonably well supported by Boost.Config "as is", I think the issue was how to build with the Apache lib when it's *not* the one that comes with your compiler, which I suspect is a bit trickier.
Agree, and I'm sure it's also evolving....so shouldn't we have a new config setup for it that starts from current RW config? Of course to get this right we really need someone to setup at least one test machine... Jeff
participants (5)
-
Boris Gubenko -
Daniel James -
Duft Markus -
Jeff Garland -
John Maddock