Triggered with virt-manager: new vm -> customize before install with a
non qemu. Guest.get_xml_config copies each device so it can set defaults
in a non persistent way. Unfortunately VirtualGraphics was copied before
a keymap lookup was ever done. End result was that we called
util.default_keymap over and over which is slow and floods the logs.
There a some problems with autobuild.sh and this patch addresses
them:
1. "python setup.py install" needs to use --root= instead
of --prefix=
2. For the rpmbuild, use dist/*.tar.gz instead of *.tar.gz
3. Add "export AUTOBUILD_OVERRIDE_VERSION=y" so that if
the version-id changes are applied, a simpler version-id
is used for autobuild.
4. Assume tests are fixed and enable running
"python setup.py test"
.
Signed-off-by: Gene Czarcinski <gene@czarc.net>
With no MANIFEST.in, distutils/setup/sdist gets the filelist
mostly correct except that it adds in whatever is in the
top directory. Using git-ls-files is not a good solution
either because a number of additional files (not git
managed) will be needed.
The solution is to use a MANIFEST.in. All of the
toplevel files which are to be included will need to be
specified. Also, each directory will need to be specified
and unwanted files (e.g., *.pyc) excluded. Other (non-specified)
files and directories will be ignored.
.
Signed-off-by: Gene Czarcinski <gene@czarc.net>
In order to fix last tests that fail for me, I had to modify the
output xml, both modified machines will fit in the first numa node, so
the cpus from that node should be mentioned in the xml and that lead
me to the fact that the problem was in the test xml and not the code.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
For <hostdev> element, mode="capabilities" is used only for
block/character device passthrough and thus it fails with pci
passthrough.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
After commit 590c8dbebf, the tests got
broken due to the fact that the commit broke something that was
working. I'm changing the error to warning and fixing the tests.
Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
Notice the missing - in -vcpus. It maps to
--hvm --cdrom pus
If the user also specifies --paravirt, they get an error that hvm and
paravirt conflict, which is very confusing.
There isn't a nice way to catch this issue which isn't back compatible,
so scrape the raw argv and try to figure it out.
This patch updates the tui python to use ipaddr
from python-ipaddr rather than IPy.py. This patch
removes all IPy dependency.
This patch works under RELEASE-0.9.4-1 but tui is
broken under branch gtk3.2. Applying patch anyway
and then will worry about it working later.
Signed-off-by: Gene Czarcinski <gene@czarc.net>
With this patch, basic IPv6 support has been added to
the network creation wizard. It includes changes to
both the glade-3 ui and the related python code. There
have been some significant changes made to both files for
IPv6 support and some additional capabilities.
This update includes the fix so color will work in the
network creation wizard.
Besides basic IPv6 support, this includes optionally
specifying a DNS domain name and enabling IPv6 internal
routing if no IPv6 address is specified.
.
Signed-off-by: Gene Czarcinski <gene@czarc.net>
