On Sun, 3 Feb 2002, Francois Gouget wrote: > > main gets it as constructed by the C runtime, isn't it? > > Which main? > * the main of a Unix application. The C runtime invokes main so it gets > the pointers from it after a fashion. But I am pretty sure these point > to a memory block that was initialized by the kernel itself > * the main of a Winelib application. This one gets it from Wine. More > precisely it gets it from scheduler/process.c:build_argv() > * and the WinMain of a Winelib application gets its command line from > memory/environ.c:ENV_BuildCommandLine(), which builds it from the result > of build_argv() > The main of a windows C program, I mean. Well, lcclnk.exe produces PE executables, so I guess that is what it is. Its main gets whatever lcccrt0.o sees fit to give it, I think. > > If you have a 'z:' drive that points to /, then the first path is > valid. Except if Wine changes the current drive or chooses another > current driver for some reason... For instance here /home/fgouget > corresponds to y:\ and z:\home\fgouget so I am not sure which one Wine > would actually choose on startup. And if it chooses y:\, then > /home/fgouget/foo.exe would not be a valid path anymore. > Well, /home/whit/cmdl is H:, and /gp/windose is C:, but what goes in GetCommandLine() and argv[0] is what was on the unix command line. If we wanted to try to idiot-proof the thing, maybe we would convert it to a windows path, but that way lies madness. If my damned app wants to parse the command line, I don't see why I shouldn't give it a command line it can parse. Maybe something to that effect should go into the doco. > > -- > Francois Gouget fgouget@free.fr http://fgouget.free.fr/ > The software said it requires Win95 or better, so I installed Linux. > Lawson It is impossible to make anything foolproof because fools are so ingenious. _______________________________________________ wine-users mailing list wine-users@winehq.com http://www.winehq.com/mailman/listinfo/wine-users
