http://bugs.winehq.org/show_bug.cgi?id=12425 --- Comment #13 from Janne <janne.kekkonen@gmail.com> --- Hello, I tested this with Wine 11.14 on a clean 32-bit prefix configured as Windows 2000. The Win16 test suite was compiled with OpenWatcom 2.0. The Makefile was modified from this: CFLAGS =-D_X86_ -DSTANDALONE '-i=C:\WATCOM\h\win;C:\WATCOM\h'-zW -e25 -zq -od -d2 -w4 -wcd=126 -wcd=102 -wcd=112 -bw-bt=windows to this: CFLAGS =-D_X86_ -DSTANDALONE -DNULL=0 '-i=C:\WATCOM\h\win;C:\WATCOM\h' -zW -e25 -zq -od -d2 -w4 -wcd=126 -wcd=102 -wcd=112 -bw-bt=windows Building with the modified Makefile was successful. comm16.exe was started with this command: WINETEST_PLATFORM=wine wine comm16.exe --log && ! grep ' [0-9]*[1-9] failures' *.log && touch comm16.ok A crash occurred with this line at the top of the backtrace: =>0 0x77873664 BuildCommDCB16+0x54(device="COM1:2400,n,8,1", lpdcb=00000000) [/usr/src/packages/BUILD/dlls/user.exe16/comm.c:427] in user.exe16 (0x0123fb68) The backtrace and source code show that lpDCB was/is NULL: https://github.com/austin987/win16test/blob/master/src/comm16.c#L63 In Wine, there is no check for lpDCB being NULL before calling memset: https://gitlab.winehq.org/wine/wine/-/blob/wine-11.15/dlls/user.exe16/comm.c... Would adding a check before memset resolve this? For example: if (lpdcb == NULL) { return -1; } -- Do not reply to this email, post in Bugzilla using the above URL to reply. You are receiving this mail because: You are watching all bug changes.