Hi,

On 11 May 2018 at 11:24, Martin Steigerwald <[email protected]> wrote:
> Package: btrfs-progs
> Version: 4.15.1-2
> Severity: wishlist
>
> Dear Dmitry,
>
> Today I wanted switch from LZO to ZSTD support. But I stumbled upon:
>
>  btrfs-progs (4.15.1-2) unstable; urgency=medium
>  .
>    * If libzstd is provided in both deb & udeb variants (such as Ubuntu
>      bionic) enable zstd support. If libzstd is not provided with both deb
>      & udeb variants (such as current Debian unstable) disable zstd
>      support. If and when libzstd is provided in both deb & udeb variants,
>      a binNMU of this package is sufficient to enable zstd support. Closes:
>      #886968
>
> I skimmed through that bug report, but I did not get what it will mean
> for ZSTD support in Debian.
>
> Also your changelog entry looks like it could happen never or at any time
> ("If and when").
>
> So I open this wishlist report to state that I like to see ZSTD support
> in btrfs-progs as being packaged in Debian. As it reads as tough you chose
> not be the one deciding upon it, feel free to reassign this to whoever is in
> charge (Debian installer team probably).
>
> I wonder about just compiling it myself. It should be safe, cause even when
> update-initramfs still stuffes the packaged tools into the inital ram disk,
> the kernel will be available to use zstd compression files.
>

I uploaded udeb support for libzstd, which got rejected by an FTP
Trainee without any reason given. Thus I lost interest in working on
that, hence the upload to btrfs-progs to disable zstd support.
Introducing a second compilation of btrfs, with/without zstd, for
deb/udeb respectively, is doable - for Debian only builds. And I'd be
willing to take a patch which does that, but not interested in writing
that myself.

This bug report is thus effectively duplicate of #886968 / reopening of it.

I guess I can reassign this to the libzstd package?!

Ideally, kernel, btrfs-progs, and d-i would all support the same
features set, like it is the case in Ubuntu...

Regards,

Dimitri.

Reply via email to