Hi,
I was planning to send a note about all this anyway. I also had
to make one other small change, to TVirtualPad.cxx:
//______________________________________________________________________________
// James Peachey 2003/03/07 for ISDC made the following change:
// The extra parentheses in the gPad macro definition were confusing
// Sun Workshop 7. Since this file is compiled (not used by cint)
// replacing the macro by its contents as defined in TVirtualPad.h,
// except without the extra parentheses.
// TVirtualPad *&gPad
TVirtualPad *&TVirtualPad::Pad()
// End of change on behalf of ISDC.
I don't understand why either piece of code caused problems,
nor did I communicate with Sun about either issue, so phrases
like "compiler bug" are my own characterizations.
If anyone has any questions about this, or sees better hacks
to make, please let me know.
Regards,
James
On Mon, 17 Mar 2003, Bruce O'Neel wrote:
> Hi,
>
> James found this (and a big thanks for that!) Here is his
> explanation from the code:
>
> // James Peachey 2003/03/10 for ISDC, made the following
> // change from the original Root 3.04.02 distribution to allow
> // compilation with Solaris Workshop 7. A compiler bug prevents
> // the original version from compiling due to the (LEGAL!) use of
> // STL objects inside a function with C linkage. The function body
> // of the original function:
> // extern "C" const char* G__saveconststring(const char* s)
> // was moved into a static function with C++ linkage, and the
> // original function simply calls this new function.
>
> // New static function with C++ linkage:
> // extern "C" const char* G__saveconststring(const char* s)
> static const char* static_G__saveconststring(const char* s)
> {
> static set<string> conststring;
> string str(s);
> conststring.insert(string(str));
> set<string>::iterator p = conststring.lower_bound(str);
> return((*p).c_str());
> }
>
> // New behavior: call static function above.
> extern "C" const char* G__saveconststring(const char* s)
> { return static_G__saveconststring(s); }
> // End of change made on behalf of ISDC.
>
>
> I'm sure that James can fill you in more. Once this is done Sun
> Workshop 7 builds and runs root just fine (as far as I have tried).
>
> cheers
>
> bruce
>
>
> Aksel Hallin writes:
> > I have the identical problem with CC on sun solaris and sun CC 5.3
> >
> > The compiler version is
> > CC: Sun WorkShop 6 update 2 C++ 5.3 Patch 111685-12 2002/12/16
> >
> > CC -O -KPIC -DG__REGEXP1 -DG__UNIX -DG__OSFDLL -Iinclude -DG__SHAREDLIB
> > -DG__ROOT -DG__REDIRECTIO -ptr/home/hpc1089/root3/root -o
> > cint/src/Api.o -c cint/src/Api.cxx
> > "/opt/SUNWspro/WS6U2/include/CC/Cstd/./functional", line 162: Error: The
> > base "std::binary_function<std::T, std::T, bool>" must be a previously
> > defined class or struct.
> > "/opt/SUNWspro/WS6U2/include/CC/Cstd/./set", line 82: Where: While
> > specializing "std::less<std::T>".
> > "/opt/SUNWspro/WS6U2/include/CC/Cstd/./set", line 82: Where:
> > Specialized in non-template code.
> > 1 Error(s) detected.
> >
> > I was trying to install root 3.05.02; however, this problem occurred
> > when the compiler was "upgraded"; before that root installed without
> > problems.
> >
> > The fix of
> > #include <functional>
> > seems to have no effect.
> >
> > Are there any other suggestions?
> >
> > Thanks,
> > Aksel Hallin
> > Queen's/SNO
> >
> >
> > > -----Original Message-----
> > > From: owner-roottalk@pcroot.cern.ch
> > [mailto:owner-roottalk@pcroot.cern.ch] On Behalf Of Philippe
> > > Canal
> > > Sent: Thursday, March 06, 2003 9:24 AM
> > > To: bruce.oneel@obs.unige.ch; roottalk@pcroot.cern.ch
> > > Subject: RE: [ROOT] Root with Sun Workshop 7
> > >
> > > Hi,
> > >
> > > In Api.cxx you could try to add the line:
> > > #include <functional>
> > > before the line
> > > #include <set>
> > >
> > > Cheers,
> > > Philippe.
> > >
> > > -----Original Message-----
> > > From: owner-roottalk@pcroot.cern.ch
> > > [mailto:owner-roottalk@pcroot.cern.ch]On Behalf Of Bruce O'Neel
> > > Sent: Thursday, March 06, 2003 6:41 AM
> > > To: roottalk@pcroot.cern.ch
> > > Subject: [ROOT] Root with Sun Workshop 7
> > >
> > >
> > > Hi,
> > >
> > > Since we're getting close to a software release we're trying to get
> > > away from 3+ year old compilers. Ergo we just upgraded to Sun
> > > Workshop 7 (or what ever it is they call it this week).
> > >
> > > This comes with Sun C++ 5.4 which seems not to build trying to
> > > compile:
> > >
> > > CC -O -KPIC -DG__REGEXP1 -DG__UNIX -DG__OSFDLL -Iinclude
> > -DG__EXCEPTION -DG
> > > __ST
> > > D_EXCEPTION -DG__SHAREDLIB -DG__ROOT -DG__REDIRECTIO
> > -ptr/home/isdc/oneel/pc
> > > 7-sc
> > > ratch/src/sparc-sun-solaris2.8 -o cint/src/Api.o -c cint/src/Api.cxx
> > > "/isdc/sw/compilers/WS/7/SUNWspro/prod/include/CC/Cstd/./functional",
> > line
> > > 162:
> > > Error: The base "std::binary_function<std::T, std::T, bool>" must be a
> > > previousl
> > > y defined class or struct.
> > > "/isdc/sw/compilers/WS/7/SUNWspro/prod/include/CC/Cstd/./set", line
> > 82:
> > > Wher
> > > e: While specializing "std::less<std::T>".
> > > "/isdc/sw/compilers/WS/7/SUNWspro/prod/include/CC/Cstd/./set", line
> > 82:
> > > Wher
> > > e: Specialized in non-template code.
> > > 1 Error(s) detected.
> > >
> > >
> > > Am I just banging my head against the wall or is this solvable?
> > >
> > > Thanks!
> > >
> > > cheers
> > >
> > > bruce "who's just about to learn much more about templates then he
> > > knew before"
> > >
> > > --
> > > Reality is 80m polygons - Alvy Ray Smith
> > >
> > > Bruce O'Neel phone: +41 22 950 91 57
> > > INTEGRAL Science Data Centre +41 22 950 91 00 (switchb.)
> > > Chemin d'Ecogia 16 fax: +41 22 950 91 35
> > > CH-1290 VERSOIX e-mail: Bruce.Oneel@obs.unige.ch
> > > Switzerland WWW: http://isdc.unige.ch/
> >
>
> --
> Reality is 80m polygons - Alvy Ray Smith
>
> Bruce O'Neel phone: +41 22 950 91 57
> INTEGRAL Science Data Centre +41 22 950 91 00 (switchb.)
> Chemin d'Ecogia 16 fax: +41 22 950 91 35
> CH-1290 VERSOIX e-mail: Bruce.Oneel@obs.unige.ch
> Switzerland WWW: http://isdc.unige.ch/
>
This archive was generated by hypermail 2b29 : Thu Jan 01 2004 - 17:50:10 MET