Thx a lot for the url. Indeed, it solves my problem with PM_Move (and PM_xxx in fact).
But now, hlds asks for "VecBModelOrigin__FP9entvars_s" again :( I've tried to define the "mangled function" it with extern "C" (yep, crappy hack :( ), but it didn't find... What's the matter again ? Micha�l "Cortex" Monerau. cort at meloo dot com == C++ power ! == Pat Magnan wrote: > http://www.thewavelength.net/forums/index.php?act=ST&f=13&t=8382&s=5a32ae437 13bc6b242369bbb911d2af1 > > That thread may be helpful, you're having similar problems, using > g++, you > get into name mangling issues. I also think since you're copying libs > about, something strange happened when you installed a > compiler/library. > The correct symlinks to libblah.so.x are setup if things are installed > correctly, or your ld.so.conf is not updated properly. > > PM_Move is C code, so compiling with g++, you get a mangled name, and > the > engine won't find it. > > At 05:00 PM 6/20/2003 +0200, you wrote: >> mmh :( >> Doesn't work either... HLDS asks for : >> "VecBModelOrigin__FP9entvars_s". >> >> Please !! Why do I have so many problems with linux :p ? >> >> Cortex. >> >> D. Hofer wrote: >>> For me it works only when i compile it with gcc and link it with >>> g++... so >>> set gcc as the compiler, run the makefile then change the compiler >>> to >>> g++ (in the makefile) and run it again. >>> >>> cya Spin >>> >>> On Fri, 20 Jun 2003 15:05:10 +0200, Cortex <[EMAIL PROTECTED]> wrote: >>> >>>> Hi, >>>> >>>> The solution of Botman didn't work (-lstdc++). >>>> >>>> I tried the brut way... I have copied the libstdc++[bla].so.2 to my >>>> /lib directory, and now, hlds finds it. >>>> >>>> However, now, it doesn't find HL SDK symbols :( AAAARGGG >>>> >>>> If I compile with g++, it doesn't find "PM_Move". >>>> If I compile with egcc, it doesn't find BModelSize (or something >>>> like >>>> that, >>>> with a horrible mangled name). HLDS doesn't seem to find the >>>> mangled >>>> name... >>>> Have I to change the mangling name version of the compilator ? >>>> >>>> (Rest of the answer below). >>>> >>>> Florian Zschocke wrote: >>>>>> I think my problem comes from the fact I use _snprintf and some >>>>>> STL... >>>>> >>>>> What is _snprintf? Usually that _something is M$ stuff (like >>>>> _strnicmp(). God knows why they have to prefix them with an >>>>> underscore). Using STL should be no problem, we use deques from >>>>> the >>>>> STL in Admin Mod >>>>> and can link with libstd++ fine. >>>> >>>> _snprintf is a function I used to remplace ALL the sprintf in the >>>> SDK... >>>> Indeed, I had a memory problem and thought it was a buffer overrun >>>> caused >>>> by >>>> a too long sprintf. It didn't come from sprintf, but I let the >>>> _snprintf. >>>> >>>> But now, this problem is resolved if I place the libstdc++ in /lib. >>>> >>>>>> When >>>>>> I put the "-static" option instead of "-shared" and when I >>>>>> compile >>>>>> with egcc, it doesn't find these symbols... >>>>> >>>>> Uhm, you don't want to use -static. YOu want to produce a dynamic >>>>> shared object, a .so, so you will need to link with -shared which >>>>> tell the >>>>> linker that you want a DSO created. >>>> >>>> OK. >>>> >>>>>> So, how could I link statically to libstdc++ so that the .so >>>>>> "contains" the libstdc++[bla].so ? >>>>> >>>>> You cannot. The libstdc++[bla].so is just that, a DSO. You cannot >>>>> link statically against a DSO. If you want to link statically >>>>> against a >>>>> library you need to use the static version of the library, the >>>>> archive. >>>>> You need to link against libstdc++[bla].a <-- >>>> >>>> OK, I think I'll let the libstdc++ thing, and distribute >>>> libstdc++[Bla].so.2 >>>> with my mod... Crappy, but if it solves the pb, it's good ;) >>>> >>>> Thx for your helps >>>> >>>> Cortex. >>>> >>>> >>>> _______________________________________________ >>>> To unsubscribe, edit your list preferences, or view the list >>>> archives, >>>> please visit: >>>> http://list.valvesoftware.com/mailman/listinfo/hlcoders >>>> >>>> >>>> >>> >>> >>> >>> _______________________________________________ >>> To unsubscribe, edit your list preferences, or view the list >>> archives, please visit: >>> http://list.valvesoftware.com/mailman/listinfo/hlcoders >> >> _______________________________________________ >> To unsubscribe, edit your list preferences, or view the list >> archives, >> please visit: >> http://list.valvesoftware.com/mailman/listinfo/hlcoders > > > _______________________________________________ > To unsubscribe, edit your list preferences, or view the list > archives, please visit: > http://list.valvesoftware.com/mailman/listinfo/hlcoders _______________________________________________ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlcoders
