Ralf, I think the problem is that TestFunc.o remains after you make C version DLL. When you try to build C++ version, it confuses that TestFunc.o is already built. Do 'make -f Makefile clean' before you make C++ version. It should work fine. $ makecint -p -mk Makefile -dl testfunc.dl -h TestFunc.h -C TestFunc.c $ make -f Makefile clean $ make -f Makefile $ makecint -p -mk Makefile -dl TestFunc.dl -H TestFunc.h -C++ TestFunc.C $ make -f Makefile clean $ make -f Makefile About garbage in /tmp directory, Cint is supposed to delete them when program termination. It is done in source file src/end.c line 591. Temp files are left only when cint is terminated abnormally. I could not find any clean solution to this. If temp files always remain, there must be some problem. Masaharu Goto > Hello Masa, > > (my last mail bounced - maybe the mail system feels better today) > > being pleased with your recent development of cint, I unfortunately have > now run into some trouble regarding the assignment of struct variables. > > The cint version 5.13.51 gives me a sigsegv when assigning a struct > returned by a C++ function compiled as dl library. The assignment works > when using an identical C function. > > (System configuration: Sun SPARC, Solaris 2.5.1 aka SunOS 5.5.1, gcc > 2.7.2.3.f.1) > > After generating the dl libraries by using > > (C program) > makecint -p -mk Makefile -dl testfunc.dl -h TestFunc.h -C TestFunc.c > > (C++ program) > makecint -p -mk Makefile -dl TestFunc.dl -H TestFunc.h -C++ TestFunc.C > > I use the following code (loaded as .C or .c file) for assigning structs: > > Could you please point me towards the right direction? > > And maybe another question: my /tmp directory gets flooded by temporal > files (preprocessor output, option '-p') left behind by cint after > building a dl library. Does there exist some way (e.g. a 'G__'-define) to > remove these files automatically? - Unfortunately, I haven't found such a > way so far. > > Thank you, > > Ralf > -- > > Ralf Garionis garionis@LS11.informatik.uni-dortmund.de > > University of Dortmund Dept. of Computer Science > Germany Systems Analysis Research Group >
This archive was generated by hypermail 2b29 : Tue Jan 04 2000 - 00:34:32 MET