Earl Chew wrote:
The call to GetFileSize() doesn't match the Microsoft documentation,
and invariably mis-sizes the disk image as zero.
A couple of comments regarding your recent changes:
a. GetFileSize() returns the bits 32-63 of the file size in
the 2nd parameter. I think code like this is more better:
DWORD high_filesize;
filesize = GetFileSize(dev, &high_filesize);
if (filesize != INVALID_FILE_SIZE || GetLastError() != NO_ERROR)
*retblocks (((unsigned long long) high_filesize << 32)
+ filesize) / blocksize;
b. Use of CreateFile() causes problems because win32 doesn't know
about the cygwin mount points. Code like this is more flexible:
fd = open(file, O_RDONLY | O_BINARY);
if (fd == -1)
return EBADF;
dev = (HANDLE) get_osfhandle(fd);
...
close(fd);
c. Similarly, code in unix_io.c needs to force binary mode to
avoid getting incorrect results when cygwin decides the image
is text!
open_flags = (flags & IO_FLAG_RW) ? O_RDWR : O_RDONLY;
#ifdef O_BINARY
open_flags |= O_BINARY;
#endif
Earl
--
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/
