On Mon, 2011-11-07 at 16:35:32 +0100, Philipp Kern wrote:
> Package: dpkg
> Version: 1.16.1.1
> Severity: normal

> I wanted all udebs to use the xz compression level 0e, which is accomplished 
> by
> passing -0 -e or just -0e to xz.  However dpkg-deb currently wants an integral
> value for the compression level.  As stated in [1] -e is actually very useful
> at least for this set of packages, going from 50M (gzip/amd64) to 40M instead
> of 43M without it.  That's of course at the expense of taking longer to
> compile, but that's ok with udebs.
> 
> The differing compression of udebs will be handled in debhelper's dh_builddeb,
> which already special-cases them.  Can we get -e support in due course in
> dpkg/unstable?

Hmm, this has several issues. The first of which is that currently
-z 0 is equivalent to -Z none, so modifying that is an interface change.
In addition gzip and bzip2 do not accept a -0 value on the command line
tools, and when using their library interfaces 0 is either equivalent to
no compression (for gzip) or normalized to 1 (for bzip2).

So, accepting 0 and extreme would need to be compressor specific, and
would also need to be validated by the specific compressor, while not
insurmountable problems, the first would be annoying as an interface,
the second slightly cumbersome in the implementation and validation
time. For gzip there's the Huffman (h) and run-length (R) encoding
options, bzip2 does not have any special encoding options.

A generic option could be to accept something like the common
"fast"/"best", although for xz the first would map to either 0 or 0e,
so there would be no way to specify either extreme or no extreme in
case it was found to be unsatisfactory for some data. This also seems
annoying as an interface.

regards,
guillem



-- 
To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]

Reply via email to