[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#553707: RFH: lzma -- future of Debian squashfs-lzma



[Cc-ing debian-live mailinglist, not sure whether they are
interested in LZMA but if so they should be aware of it.]

* Jonathan Nieder <jrnieder@gmail.com> [20091103 00:14]:

> I was thinking about how to bring the lzma package more up to date,
> but I’m scared to do anything for fear of breaking squashfs-lzma. :)
> So I thought I’d write for advice.

> squashfs is a read-only compressed filesystem.  Currently a kernel
> module for reading LZMA-compressed squashfs 3.x is packaged for Debian
> through the lzma source package.  I think the userspace support is in
> squashfs-tools, though I’m not sure since I’ve never tried it.

> Good news: squashfs maintainer Phillip Lougher is pushing for
> LZMA support in squashfs 4.0 in the mainline kernel v2.6.33 or so. [1]
> Packaging this for Debian would perhaps mean backporting the change
> for the linux-2.6 package and making sure the squashfs-tools package
> has the appropriate support.

>  1. How to support current users until an updated kernel enters sid?
>     Should the lzma package have to continue to produce an lzma-source
>     package?

>     I am not excited about making sure this still works with each LZMA
>     SDK upgrade.  squashfs-lzma upstream used to just pick one version
>     to support (now the version in the mainline kernel is used), and
>     I’m afraid having to rebase the patch for each release would slow
>     things down a lot.

Assuming Squeeze will be released with kernel 2.6.32 and
squashfs/lzma will reach mainline not before 2.6.33 it might be
worth the effort to provide an according squashfs 4/lzma +
squashfs-lzma-tools toolchain.

>  2. Once squashfs 4 + lzma is available, is there a need to continue
>     to support squashfs 3 + lzma?

>     I am hoping not, because it is not clear the current kernel
>     patches apply to recent kernels.  But if there is, we can probably
>     find some way.

I personally don't think it's necessary to provide long-term support
for squashfs 3 + lzma once squashfs 4 + lzma is available, though
I'm not sure what other people think about it.

>  3. More generally, what do people use squashfs-lzma for, and what
>     guarantees do they need in order to do it?

We at grml (http://grml.org/) use it for compressing the entire
rootfs. Thanks to better compression of lzma (compared to zlib)
squashfs-lzma allows us to provide more software on the same ISO
size, though the build time clearly increases.

If you're interested in some stats:

  http://grml.org/grml-live/#lzma-vs-zlib

> I am hoping some squashfs-lzma user can explain how and perhaps take
> on the task of assuring it is well supported for squeeze.  But there
> is plenty to do short of that, and I would be glad to help your
> efforts in any way I can.

Thanks for your efforts.

JFYI: We (the grml team) have working squashfs 4 + lzma patches for
kernel 2.6.31 and an according squashfs-tools package providing
lzma support. Please let me know if we can assist in any way.

regards,
-mika-
-- 
http://michael-prokop.at/  || http://adminzen.org/
http://grml-solutions.com/ || http://grml.org/

Attachment: signature.asc
Description: Digital signature


Reply to: