Since this sounds like a question about how the Mingw version (-mno-cygwin) of the compiler works, you should probably direct the question to that list ([EMAIL PROTECTED]). Good luck, Larry Hall [EMAIL PROTECTED] RFK Partners, Inc. http://www.rfk.com 118 Washington Street (508) 893-9779 - RFK Office Holliston, MA 01746 (508) 893-9889 - FAX At 04:16 AM 12/15/2000, David Ponce wrote: >Hello, > >I am a NT Emacs 21 pretester and I successfully compiled it (with >-mno-cygwin) using the latest (great!) cygwin 1.1.6 (I tried the >20001213 snapshot too). Emacs works well except a problem using the >O_EXCL flag with the _open function. With this flag _open always >return -1 with errno=22. > >Following is a simple test program that isolates the problem: > >------------------------ >#include <io.h> >#include <stdio.h> >#include <errno.h> >#include <sys/stat.h> >#include <fcntl.h> > >int fdesc; > >int main(void) >{ > > fprintf(stderr, "O_BINARY=%i, _O_BINARY=%i\n", O_BINARY, _O_BINARY); > > fprintf(stderr, "O_WRONLY=%i, _O_WRONLY=%i\n", O_WRONLY, _O_WRONLY); > fprintf(stderr, "O_CREAT=%i, _O_CREAT=%i\n", O_CREAT, _O_CREAT); > fprintf(stderr, "O_EXCL=%i, _O_EXCL=%i\n", O_EXCL, _O_EXCL); > fprintf(stderr, "O_TRUNC=%i, _O_TRUNC=%i\n", O_TRUNC, _O_TRUNC); > fprintf(stderr, "O_NOINHERIT=%i, _O_NOINHERIT=%i\n", > O_NOINHERIT, _O_NOINHERIT); > fprintf(stderr, "S_IREAD=%i, _S_IREAD=%i\n", S_IREAD, _S_IREAD); > fprintf(stderr, "S_IWRITE=%i, _S_IWRITE=%i\n", S_IWRITE, _S_IWRITE); > > errno = 0; > fdesc = _open("testopen.tmp1", > O_BINARY | O_WRONLY | O_CREAT | O_TRUNC | O_EXCL | _O_NOINHERIT, > S_IREAD | S_IWRITE); > fprintf(stderr, "fdesc=%i\n", fdesc); > fprintf(stderr, "errno=%i\n", errno); > > errno = 0; > fdesc = _open("testopen.tmp2", > O_BINARY | O_WRONLY | O_CREAT | O_TRUNC | _O_NOINHERIT, > S_IREAD | S_IWRITE); > fprintf(stderr, "fdesc=%i\n", fdesc); > fprintf(stderr, "errno=%i\n", errno); > >} >------------------------ > >And the transcription of the test case (the testopen.tmp files does not >initially exist) > >------------------------ >C:\cygwin\tmp>gcc -mno-cygwin testopen.c > >C:\cygwin\tmp>a.exe >O_BINARY=32768, _O_BINARY=32768 >O_WRONLY=1, _O_WRONLY=1 >O_CREAT=256, _O_CREAT=256 >O_EXCL=1024, _O_EXCL=1024 >O_TRUNC=512, _O_TRUNC=512 >O_NOINHERIT=128, _O_NOINHERIT=128 >S_IREAD=256, _S_IREAD=256 >S_IWRITE=128, _S_IWRITE=128 >fdesc=-1 >errno=22 >fdesc=3 >errno=0 > >C:\cygwin\tmp>del testopen.tmp* > >C:\cygwin\tmp>gcc testopen.c > >C:\cygwin\tmp>a.exe >O_BINARY=65536, _O_BINARY=65536 >O_WRONLY=1, _O_WRONLY=1 >O_CREAT=512, _O_CREAT=512 >O_EXCL=2048, _O_EXCL=2048 >O_TRUNC=1024, _O_TRUNC=1024 >O_NOINHERIT=262144, _O_NOINHERIT=262144 >S_IREAD=256, _S_IREAD=256 >S_IWRITE=128, _S_IWRITE=128 >fdesc=3 >errno=0 >fdesc=4 >errno=0 >------------------------ > >When I compile the same program without -mno-cygwin it works well. > >So is the O_EXCL flag supported by the mingw version of _open. Or is >it a bug? > >Thank you so much for any help. > >Sincerely, >David > >-- >Want to unsubscribe from this list? >Send a message to [EMAIL PROTECTED] -- Want to unsubscribe from this list? Send a message to [EMAIL PROTECTED]
Re: _open problem using O_EXCL flag and -mno-cygwin
Larry Hall (RFK Partners, Inc) Fri, 15 Dec 2000 07:51:36 -0800
- _open problem using O_EXCL flag and -mno-cy... David Ponce
- Re: _open problem using O_EXCL flag an... Larry Hall (RFK Partners, Inc)
- Re: _open problem using O_EXCL flag an... Earnie Boyd
- Re: _open problem using O_EXCL fla... Christopher Faylor
- Re: Re: _open problem using O_EXCL fla... David Ponce
- Re: _open problem using O_EXCL flag an... Earnie Boyd
- Re: Re: _open problem using O_EXCL fla... Earnie Boyd
