From: hgfernan AT usp DOT br (Hilton Fernandes) Subject: RE: Console-free EXE with Mingw32 10 Apr 1998 13:36:50 -0700 Sender: root AT cygnus DOT com Message-ID: <9804091443.AA44566.cygnus.gnu-win32@spider.uspnet.usp.br> References: <003801bd6294$c6a313e0$fa173185 AT gbird0 DOT fu DOT is DOT saga-u DOT ac DOT jp> To: "Colin Peters" , "GNU-Win32" Cc: Hilton Fernandes MIME-Version: 1.0 Content-Transfer-Encoding: 7BIT Content-Type: text/plain; charset=US-ASCII On 8 Apr 98 at 11:19, Colin Peters wrote: > Hilton Fernandes wrote: > > To create a console executable you use, say: > > gcc -o some_exec some_obj_1.o some_obj_2.o > > > > To create a console-free, GUI executable, you append > > -W1,--subsystem,windows,-e,_mainCRTStartup > > > > I believe those "-W1" options above should be -Wl (that is, dash double-yew > ell, not dash double-yew one). Also, if you use gcc to do your linking (as > suggested) and if your specs file is correctly set up (as it should be if > everything else is working) then you could use -mwindows to do the same > thing (and get those extra libraries too) e.g. > > gcc -o hello.exe foo.o bar.o -mwindows > > If you are using 0.1.4 of Mingw32 instead of the latest snapshot (the > versions included with Jan-Jaap's versions of GCC or Mumit Khan's versions > of EGCS also use -mwindows) then you should use -windows, though I suggest > you edit the specs file and replace all occurences of %{windows: with > %{mwindows: and use -mwindows to be consistent with Cygnus and later Mingw32 > releases on this. > > Also, not specifying the entry point with -e _mainCRTStartup should not give > you a warning with the latest snapshot of Mingw32 (or version 0.1.4 either > if I remember correctly). Even if you do get the warning (could not find > entry point _WinMainCRTStartup or something like that) it is harmless in all > the cases I am aware of. > Colin Peters, Thanks for your correction. I'm sorry that i mistyped the "ell" in the "-Wl" option. I know the _WinMainCRTStartup warning is harmless, but my C/C++ programmer experience tells me to avoid any warning: between a long list of harmless warnings can be a __harmful__ warning that will make the program crash. I believe this is also the original poster intention. OTOH, i'm happy that the new compiler versions by you, Mummit Khan and Jan-Jaap are able to correct this minor nuisance. Many thanks you for your work, ++Hilton ------------------- Hilton Fernandes email: hgfernan AT usp DOT br www: http://www.lsi.usp.br/~hilton.html (inactive) M. Sc. Student of Parallel Distributed Applications at Escola Politecnica (Polytechnic School) University of S. Paulo - Brazil - For help on using this list (especially unsubscribing), send a message to "gnu-win32-request AT cygnus DOT com" with one line of text: "help".