Daniel Veillard wrote:
On Wed, Sep 09, 2009 at 12:08:13PM +0200, Jim Meyering wrote:
Daniel Veillard wrote:
On Wed, Sep 09, 2009 at 11:57:40AM +0200, Jim Meyering wrote:
Daniel Veillard wrote:
Hum, I realize that support of LZOP was added after 0.7.0, so we never made a release with it (well except for git snapshot which may have been pushed). I wonder if the best is not to just drop the lzop option altogether and stick xz as a package dependancy until we have found a way to provide at the API level which compression options are actually available.
Opinions ?
Dropping lzop sounds good. It seems lzop is not very popular. We don't need that many choices.
Maybe even nuke lzma too before we're stuck with it forever. Technically, we can do that, since it was added only a month ago, also after 0.7.0:
v0.7.0-35-g2d6a581
Actually om my machine here lzma is provided as a backward compat option by xz, so yes I'm inclined to remove that option too:
aphio:~ -> which lzma /usr/bin/lzma paphio:~ -> rpm -qf /usr/bin/lzma xz-lzma-compat-4.999.8-0.8.beta.20090817git.fc11.x86_64
I will post a patch later,
FYI, xz can decompress lzma-compressed input:
$ echo foooo |lzma -c|xz -dc foooo
so libvirt won't need the "lzma" command, as long as xz is available.
Okay, I suggest the following patch removing the 2 extra compressors and making sure the package including the daemon, if compiled with qemu has the proper dependancies. xz package dependancy is IMHO a small price to pay: Size: 443012 to garantee reliability and bzip2 and gzip are so standard that it should not be a problem for anybody to list them explicitely.
The changes to the C code look fine to me. However seeing the added dependency on xz makes me wish for a "strongly-recommended" sort of tag, rather than a hard dependency on a tool that one may never opt to use. ACK, in spite of that.