Re: autoconf-2.61's AC_LINK_IFELSE with MinGW cross-compilers

[Eric Blake]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

According to Chris Johns on 4/1/2007 11:57 PM:
> 
>> I understand that MSYS
>> has a problem, but I don't know that Cygwin does.  Does Cygwin have
>> the problem?  If not, then the test is incorrect.
> 
> It may have the problem if the file-system is FAT32 rather than NTFS. An
> RTEMS user reported Cygwin worked once he changed from FAT32 to NTFS.

On NTFS, Cygwin follows Linux and POSIX semantics.  On FAT32, the
underlying filesystem lacks mode bits, so there is no way to accurately
represent executable bits, so my understanding is that cygwin must fake
'test -x' by looking at extensions and she-bangs for regular files.  But
even then, it correctly marks FAT32 directories as searchable.  But FAT32
is inherently insecure, and the cygwin folks recommend against using it
(for example, the cygwin package of git will not work on FAT32,
intentionally choosing to require proper permissions provided by NTFS;
even though the upstream git maintainers have added a hack that blindly
bypasses file modes on cygwin on all filesystems because of the chance
that upstream maintainers felt that supporting FAT32 was worth their
while).  For the purposes of developing programs (and thus using
autoconf), I would hope that the developer is smart enough to realize the
limitations of FAT32 and develop on NTFS instead; but it does mean that
this autoconf test needs to actually test the feature of the current file
system rather than assuming global platform behavior.

- --
Don't work too hard, make some time for fun as well!

Eric Blake             address@hidden
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (Cygwin)
Comment: Public key at home.comcast.net/~ericblake/eblake.gpg
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFGEPLP84KuGfSFAYARAi8DAJ95ZhHynZi6n/h6NF+dzacnBq0YEACgoiZ4
xlQOVOXdv8ofSPK/MSA7V6E=
=8Nai
-----END PGP SIGNATURE-----