On Tue, Mar 25, 2014 at 09:03:32AM +0100, Vratislav Podzimek wrote:
On Mon, 2014-03-24 at 09:06 -0700, Brian C. Lane wrote:
> + # If this looks like a tarfile, skip trying to mount it
> + if not any(self.data.method.url.endswith(suffix) for suffix in
TAR_SUFFIX):
What about definining a 'using_tar' property and using it here and in
the install method?
Sure, I can do that.
> + # Use the archive's size for want of something
better
> + self.source_size = os.stat(self.image_path)[stat.ST_SIZE] * 2
Why is the ST_SIZE doubled? If it's intentional, it deserves a comment.
There is no way to know exactly what the uncompressed size will be, so I
guess that compression will be around 50% -- I'll expand a comment a
bit.
> +
> + self.pct_lock = Lock()
> + self.pct = 0
> + threadMgr.add(AnacondaThread(name=THREAD_LIVE_PROGRESS,
> + target=self.progress))
> +
> + cmd = "tar"
> + # preserve: permissions, owners, groups, ACL's, xattrs, times,
> + # symlinks, hardlinks
> + # go recursively, include devices and special files, don't cross
> + # file system boundaries
This comment needs a minor update when used here with tar instead of
rsync in the live install payload, I think.
Good point, I'll clean that up.
If you want to make these installations a lot faster, you can use
pbzcat/xzcat -T/pigz with their output piped into the tar itself.
However, that could be something extra for a follow-up patch.
Yeah, I was planning on doing that in a future patch.
--
Brian C. Lane | Anaconda Team | IRC: bcl #anaconda | Port Orchard, WA (PST8PDT)