On Tue, 15 Jun 2004, John Cooper wrote:
> > The point is that it's not about cygwin-vs-windoze apps. It's about
> > apps-that-use-console-stdin-and-stdout vs. apps-that-display-a-gui; those
> > that show a gui could usefully be detached, but those that read their input
> > from stdin will break if the shell detaches them.
Hi John,
I'm the maintainer for zsh on Cygwin.
> Yes, you're right, the old "native" zsh option was specifically to do with GUI
> apps rather than "Windows" apps per se - here's the doc to for enabling the
> option (it was off by default):
>
> winntwaitforguiapps: When set, makes the shell wait for win32 GUI apps to
> terminate instead of spawning them asynchronously.
>
> > I don't think there's a reliable enough mechanism by which a shell could
> > detect one case from the other.
>
> Below is the code it used to determine if a program is a GUI program or not. I
> don't know how well it works under all conditions; however it did work fine for
> me.
>
> Even if not perfectly reliable, could something like this be added but disabled
> by default? I for one would find it useful.
I guess I don't really have much of a problem with adding such a feature,
provided it's something that many users really want. I can see some
merit to it, but is it really that much work to type '&' after the
command to run it in the background?
Anyway, can you point me to where you got this code example? I want to
see where it was originally called in zsh code and possibly adapt it to
the current code base. It'll take a bit of re-writing (which I'm willing
to do) because this code makes liberal usage of win32 calls and
defines/structs, which generally aren't available/desirable to use in the
Cygwin runtime. If it works well enough, though, I might submit it for
inclusion in the zsh base.
> --- John
>
>
> /*
> How To Determine Whether an Application is Console or GUI [win32sdk]
> ID: Q90493 CREATED: 15-OCT-1992 MODIFIED: 16-DEC-1996
> */
> #include <winnt.h>
> #define xmalloc(s) HeapAlloc(GetProcessHeap(),HEAP_ZERO_MEMORY,(s))
> #define xfree(p) HeapFree(GetProcessHeap(),0,(p))
> #define XFER_BUFFER_SIZE 2048
>
> int is_gui(char *exename) {
>
> HANDLE hImage;
>
> DWORD bytes;
> DWORD SectionOffset;
> DWORD CoffHeaderOffset;
> DWORD MoreDosHeader[16];
>
> ULONG ntSignature;
>
> IMAGE_DOS_HEADER image_dos_header;
> IMAGE_FILE_HEADER image_file_header;
> IMAGE_OPTIONAL_HEADER image_optional_header;
>
>
> hImage = CreateFile(exename, GENERIC_READ, FILE_SHARE_READ, NULL,
> OPEN_EXISTING, FILE_ATTRIBUTE_NORMAL, NULL);
>
> if (INVALID_HANDLE_VALUE == hImage) {
> return 0;
> }
>
> /*
> * Read the MS-DOS image header.
> */
> if (!ReadFile(hImage, &image_dos_header, sizeof(IMAGE_DOS_HEADER),
> &bytes,NULL)){
> CloseHandle(hImage);
> return 0;
> }
>
> if (IMAGE_DOS_SIGNATURE != image_dos_header.e_magic) {
> CloseHandle(hImage);
> return 0;
> }
>
> /*
> * Read more MS-DOS header. */
> if (!ReadFile(hImage, MoreDosHeader, sizeof(MoreDosHeader),
> &bytes,NULL)){
> CloseHandle(hImage);
> return 0;
> }
>
> /*
> * Get actual COFF header.
> */
> CoffHeaderOffset = SetFilePointer(hImage, image_dos_header.e_lfanew,
> NULL,FILE_BEGIN);
>
> if (CoffHeaderOffset == (DWORD) -1){
> CloseHandle(hImage);
> return 0;
> }
>
> CoffHeaderOffset += sizeof(ULONG);
>
> if (!ReadFile (hImage, &ntSignature, sizeof(ULONG),
> &bytes,NULL)){
> CloseHandle(hImage);
> return 0;
> }
>
> if (IMAGE_NT_SIGNATURE != ntSignature) {
> CloseHandle(hImage);
> return 0;
> }
>
> SectionOffset = CoffHeaderOffset + IMAGE_SIZEOF_FILE_HEADER +
> IMAGE_SIZEOF_NT_OPTIONAL_HEADER;
>
> if (!ReadFile(hImage, &image_file_header, IMAGE_SIZEOF_FILE_HEADER,
> &bytes, NULL)){
> CloseHandle(hImage);
> return 0;
> }
>
> /*
> * Read optional header.
> */
> if (!ReadFile(hImage, &image_optional_header,
> IMAGE_SIZEOF_NT_OPTIONAL_HEADER,&bytes,NULL)) {
> CloseHandle(hImage);
> return 0;
> }
>
> CloseHandle(hImage);
>
> if (image_optional_header.Subsystem ==IMAGE_SUBSYSTEM_WINDOWS_GUI)
> return 1;
> return 0;
> }
> int is_9x_gui(char *prog) {
>
> char *progpath;
> DWORD dwret;
> char *pathbuf;
> char *pext;
>
> pathbuf=xmalloc(MAX_PATH);
>
> progpath=xmalloc(MAX_PATH<<1);
>
> if (GetEnvironmentVariable("PATH",pathbuf,MAX_PATH) ==0) {
> goto failed;
> }
>
> pathbuf[MAX_PATH]=0;
>
> dwret = SearchPath(pathbuf,prog,".EXE",MAX_PATH<<1,progpath,&pext);
>
> if ( (dwret == 0) || (dwret > (MAX_PATH<<1) ) )
> goto failed;
>
> dprintf("progpath is %s\n",progpath);
> dwret = is_gui(progpath);
>
> xfree(pathbuf);
> xfree(progpath);
>
> return dwret;
>
> failed:
> xfree(pathbuf);
> xfree(progpath);
> return 0;
>
>
> }
>
>
>
> --- John
>
> -----Original Message-----
> From: Dave Korn [mailto:[EMAIL PROTECTED]
> Sent: 15 June 2004 15:40
> To: 'John Cooper'; [EMAIL PROTECTED]
> Subject: RE: Ctrl-Z fails to suspend Windows programs
>
> > -----Original Message-----
> > From: cygwin-owner On Behalf Of John Cooper
> > Sent: 15 June 2004 15:05
> > To: cygwin
> > Subject: RE: Ctrl-Z fails to suspend Windows programs
> >
> > The old native (non-cygwin) port of zsh would somehow detect if it was
> > about to exec a Windows app, and run it as a background process, thus
> > returning a zsh prompt immediately. Could something like this be
> > added to cygwin bash/zsh?
>
> AFAICS the ability is already there. Just enter "windows_app.exe &" at a bash
> shell.
>
> > This was very useful. With the cygwin zsh, I often find myself
> > invoking a Windows app and not being able to get back to the shell
> > window without first terminating the Windows app.
>
> Well, the same goes if you run a cygwin app: you don't get the prompt back until it
> exits.
>
> The point is that it's not about cygwin-vs-windoze apps. It's about
> apps-that-use-console-stdin-and-stdout vs. apps-that-display-a-gui; those that
> show a gui could usefully be detached, but those that read their input from
> stdin will break if the shell detaches them. I don't think there's a reliable
> enough mechanism by which a shell could detect one case from the other.
>
> Try starting insight (gui version of gdb) from the bash prompt. You also won't
> get a shell back until it exits.
>
>
> cheers,
> DaveK
> --
> Can't think of a witty .sigline today....
>
>
> --
> Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
> Problem reports: http://cygwin.com/problems.html
> Documentation: http://cygwin.com/docs.html
> FAQ: http://cygwin.com/faq/
>
--
Peter A. Castro <[EMAIL PROTECTED]> or <[EMAIL PROTECTED]>
"Cats are just autistic Dogs" -- Dr. Tony Attwood
--
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Problem reports: http://cygwin.com/problems.html
Documentation: http://cygwin.com/docs.html
FAQ: http://cygwin.com/faq/
