Re: [PATCH] Allow ext4 without magic argument (keep a flag for migrate)

[Eric Sandeen <sandeen redhat com>]

Jeremy Katz wrote:
> ext4 should be safe enough to allow people to select it without a magic flag

Looks good to me, thanks.  I need to try the migration, there were
issues that this migration path was different from others, and needed to
be done pre-install, not -post.  (i.e. tune2fs -O extents, mount as
ext4, then perform upgrade) - any idea if this has been addressed?

Thanks,
-Eric

>  fsset.py |    9 +--------
>  1 files changed, 1 insertions(+), 8 deletions(-)
> 
> diff --git a/fsset.py b/fsset.py
> index 3472b94..3972094 100644
> --- a/fsset.py
> +++ b/fsset.py
> @@ -694,7 +694,7 @@ class ext3FileSystem(extFileSystem):
>          self.name = "ext3"
>          self.extraFormatArgs = [ "-t", "ext3" ]
>          self.partedFileSystemType = parted.file_system_type_get("ext3")
> -        if flags.cmdline.has_key("ext4"):
> +        if flags.cmdline.has_key("ext4migrate"):
>              self.migratetofs = ['ext4']
>  
>      def formatDevice(self, entry, progress, chroot='/'):
> @@ -721,13 +721,6 @@ class ext4FileSystem(extFileSystem):
>          self.extraFormatArgs = [ "-t", "ext4" ]
>          self.bootable = False
>  
> -        # this is way way experimental at present...
> -        if flags.cmdline.has_key("ext4"):
> -            self.supported = -1
> -        else:
> -            self.supported = 0
> -
> -
>      def formatDevice(self, entry, progress, chroot='/'):
>          extFileSystem.formatDevice(self, entry, progress, chroot)
>          extFileSystem.setExt3Options(self, entry, progress, chroot)