Saturday, June 18, 2016

virt-manager 1.4.0 release

I've just released virt-manager 1.4.0. Besides the spice GL bits that I previously talked about, nothing too much exciting in this release except a lot of virt-install/virt-xml command line extensions.

The changelog highlights:
  • virt-manager: spice GL console support (Marc-André Lureau, Cole Robinson)
  • Bump gtk and pygobject deps to 3.14
  • virt-manager: add checkbox to forget keyring password (Pavel Hrdina)
  • cli: add --graphics gl= (Marc-André Lureau)
  • cli: add --video accel3d= (Marc-André Lureau)
  • cli: add --graphics listen=none (Marc-André Lureau)
  • cli: add --transient flag (Richard W.M. Jones)
  • cli: --features gic= support, and set a default for it (Pavel Hrdina)
  • cli: Expose --video heads, ram, vram, vgamem
  • cli: add --graphics listen=socket
  • cli: add device address.type/address.bus/...
  • cli: add --disk seclabelX.model (and .label, .relabel)
  • cli: add --cpu cellX.id (and .cpus, and .memory)
  • cli: add --network rom_bar= and rom_file=
  • cli: add --disk backing_format=

Friday, June 10, 2016

check-pylint: mini tool for running pylint anywhere

pylint and pep8 are indispensable tools for python development IMO. For projects I maintain I've long ago added a 'setup pylint' sub-command to run both commands, and I've documented this as a necessary step in the contributor guidelines.

But over the years I've accumulated many repos for small bits of python code that never have need for a setup.py script, but I still want the convenience of being able to run pylint and pep8 with a single command and a reasonable set of options.

So, a while back I wrote this tiny 'check-pylint' script which does exactly that. The main bit it adds is automatically searching the current directory for python scripts and modules and passing them to pylint/pep8. From the README:

Simple helper script that scoops up all python modules and scripts beneath the current directory, and passes them through pylint and pep8. Has a bit of smarts to ignore .git directory, and handle files that don't end in .py

The point is that you can just fire off 'check-pylint' in any directory containing python code and get a quick report.

Tuesday, June 7, 2016

python-bugzilla API changes in git

I've made a number of API changes to python-bugzilla in git. Most of it is removing old mis-designed features which I don't think anyone is actually using. I went through all apps I know of that use python-bugzilla (fedora bits like bodhi pkgdb infra scripts, bugwarrior, searchcode.com) to confirm as much and I don't think anything below will affect them. Even if there's no complaints I won't cut the release for 4-6 weeks.

The list is long but most changes are fairly straight forward. If any of this sounds controversial, or if I misjudged one of the 'unused' bits and should reconsider, please leave your thoughts here or on the python-bugzilla mailing list.

The interesting change is:
  • Bugzilla.bug_autorefresh now defaults to False. Previously if trying to access Bug.foo, and 'foo' wasn't cached, the Bug() object would basically do a Bug.refresh() and fetch all the bug contents from the server. This default sucks, it can lead to poorly performing scripts and unnecessary bugzilla load. If you have code that uses include_fields, exclude_fields, or extra_fields anywhere, this change may affect you. You can set Bugzilla().bug_autorefresh = False before doing any bug lookup to force this off (with current versions) and catch any errors. If you hit issues you probably need to extend your include_fields specifications. The reason for this change is that the old pattern made it too easy for people's scripts to unintentionally start requiring a much higher number of XMLRPC calls, thus completely negating the usage of include_fields in the first place. Some more details over here: https://github.com/python-bugzilla/python-bugzilla/blob/master/examples/bug_autorefresh.py https://bugzilla.redhat.com/show_bug.cgi?id=1297637

These are slightly interesting and may impact a few people:
  • Credentials are now cached in ~/.cache/python-bugzilla/. If the old ~/.bugzilla{cookies,token} files are in place we will continue to update and use them. * RHBugzilla.rhbz_back_compat __init__ attribute is gone. If the user manually set this to true, we would alter some Bug fields returned via query() to convert python lists into a single comma separated string, to match rhbugzilla output from before the 2012 upgrade. Convert your code to use the modern bugzilla list output.
  • bin/bugzilla was converted to argparse, which has a bug on python 2.7 that affects some possible command lines: http://bugs.python.org/issue9334 If you use a command like 'bugzilla modify --cc -foo@example.com' to remove that email from the CC list, you now need to ensure there's an equals sign in there, like --cc=-foo@example.com. That is backwards compatible with old python-bugzilla too, so update your scripts now. A few other options expect that format too.
  • bugzilla query --boolean_chart option is removed. It provided a custom specification for crafting complex queries, similar to what the bugzilla UI can do. I don't think anyone is really using this, but if you are, generate a web query URL and pass it to bugzilla query --from-url '$URL' which is much easier to deal with.
  • bugzilla query 'boolean' options, where you could say 'bugzilla query --keywords 'foo & bar' to match both substrings, are no longer supported. If you need logic like this, use the --from-url technique mentioned above.
  • Bug.get_history() is now Bug.get_history_raw() Bugzilla.bugs_history is now Bugzilla.bugs_history_raw() This API is only a year old and unwisely returns raw output from bugzilla, which has some formatting oddities. I renamed it to *_raw so we have the future opportunity to add an API with better output.
  • The getbugsimple and getbugssimple APIs were removed. They were basically just wrappers around standard getbug() at this point. Use getbug/getbugs instead.
  • The simplequery API was removed. This basically matches the basic query from the bugzilla front page. Just use the standard query methods instead, see examples/query.py from git. I didn't find any users of this function.
  • Various whiteboard editing functions from the Bug object were removed. - getwhiteboard: this is just a wrapper around standard Bug attributes - appendwhiteboard, prependwhiteboard, setwhiteboard: these were just wrappers around standard build_update/update_bugs - addtag/gettags/deltag which minor wrappers around the *whiteboard functions, and were poorly named, given that there's an actual bug 'tags' field at this point. See examples/update.py from git for examples of using the standard update APIs. That said I've never heard of anyone actually using these, and they were designed around really old RHBZ APIs.

These I doubt will actually affect anyone:
  • Bugzilla.initcookie() dropped... use Bugzilla.cookiefile = X instead 
  • Bugzilla.adduser() dropped... it was the same as Bugzilla.createuser 
  • RHBugzilla.multicall __init__ attribute is gone. It's been a no-op and raises a warning for a long time. 
  • Bugzilla.version string is gone... this was meant to describe the python-bugzilla API version but was never used. 
  • Initing like Bugzilla() with no options was previously allowed, but now requires an explicit Bugzilla(url=None). 
  • We no longer handle cookies in LWP format. We've been silently converting them to mozilla format for 2 years so I assume this doesn't affect anyone, but if it does, you might need to delete your caches in ~/.bugzilla{cookies,token} 
  • Bug.setstatus(), Bug.close() args 'private_in_it' and 'nomail' were removed: they have been no-ops for years 
  • Bug.addcomment() args 'timestamp', 'bz_gid' and 'worktime' were removed: they've been no-ops for years
  • Bug.setassignee() arg 'reporter' was removed: it's thrown an error for years

Sunday, May 22, 2016

spice OpenGL/virgl acceleration on Fedora 24

New in Fedora 24 virt is 3D accelerated SPICE graphics, via Virgl. This is kinda-sorta OpenGL passthrough from the VM up to the host machine. Much of the initial support has been around since qemu 2.5, but it's more generally accessible now that SPICE is in the mix, since that's the default display type used by virt-manager and gnome-boxes.

I'll explain below how you can test things on Fedora 24, but first let's cover the hurdles and caveats. This is far from being something that can be turned on by default and there's still serious integration issues to iron out. All of this is regarding usage with libvirt tools.

Caveats and known issues

Because of these issues, we haven't exposed this as a UI knob in any of the tools yet, to save us some redundant bug reports for the above issues from users who are just clicking a cool sounding check box :) Instead you'll need to explicitly opt in via the command line.

Testing it out

You'll need the following packages (or later) to test this:
  • qemu-2.6.0-2.fc24
  • libvirt-1.3.3.1-2.fc24
  • virt-manager-1.3.2-4.20160520git2204de62d9.fc24
  • At least F24 beta on the host
  • Fedore 24 beta in the guest. Anything earlier is not going to actually enable the 3D acceleration. I have no idea about the state of other distributions. And to make it abundantly clear this is linux only and likely will be for a long time at least, I don't know if Windows driver support is even on the radar.
Once you install a Fedora 24 VM through the standard methods, you can enable spice GL for your VM with these two commands:

$ virt-xml --connect $URI $VM_NAME --confirm --edit --video clearxml=yes,model=virtio,accel3d=yes
$ virt-xml --connect $URI $VM_NAME --confirm --edit --graphics clearxml=yes,type=spice,gl=on,listen=none

The first command will switch the graphics device to 'virtio' and enable the 3D acceleration setting. The second command will set up spice to listen locally only, and enable GL. Make sure to fully poweroff the VM afterwards for the settings to take effect. If you want to make the changes manually with 'virsh edit', the XML specifics are described in the spice GL documentation.

Once your VM has started up, you can verify that everything is working correctly by checking glxinfo output in the VM, 'virgl' should appear in the renderer string:

$ glxinfo | grep virgl
    Device: virgl (0x1010)
OpenGL renderer string: Gallium 0.4 on virgl

And of course the more fun test of giving supertuxkart a spin :)

Credit to Dave Airlie, Gerd Hoffman, and Marc-André Lureau for all the great work that got us to this point!

Wednesday, January 20, 2016

Tips for querying git tags


With package maintenance, bug triage, and email support, I often need to look at a project's git tags to know about the latest releases, when they were released, and what releases contain certain features. Here's a couple workflow tips that make my life easier.

Better git tag listing

Based on Peter Hutterer's 'git bi' alias for improved branch listing (which is great and highly recommended), I made one for improved tags output that I mapped as 'git tags'. Output looks like:

  • Shows tag name, commit message, commit ID, and date, all colorized. Commit message is redundant for many projects that tag the release commit, but it's interesting in some cases.
  • Tags are listed by date rather than alphabetically. Some projects change tag string formats, or versioning schemes, that then don't sort correctly when listed alphabetically. Sorting by date makes it easy to see the latest tag. Often I just want to know what the latest tag or the latest stable release is, this makes it easy.
The alias code is:
 [alias]  
   tags = "!sh -c ' \  
 git for-each-ref --format=\"%(refname:short)\" refs/tags | \  
 while read tag; do \  
   git --no-pager log -1 --format=format:\"$tag %at\" $tag; echo; \  
 done | \  
 sort -k 2 | cut -f 1 --delimiter=\" \" | \  
 while read tag; do \  
   fmt=\"%Cred$tag:%Cblue %s %Cgreen%h%Creset (%ai)\"; \  
   git --no-pager log -1 --format=format:\"$fmt\" $tag; echo; \  
 done'"  

Find the first tag that contains a commit

This seems to come up quite a bit for me. An example is here; a user was asking about a virt-install feature, and I wanted to tell them what version it appeared in. I grepped git log, found the commit, then ran:

 $ git describe --contains 87a611b5470d9b86bf57a71ce111fa1d41d8e2cd  
 v1.0.0~201  

That shows me that v1.0.0 was the first release with the feature they wanted, just take whatever is to the left of the tilde.

This often comes in handy with backporting as well: a developer will point me at a bug fix commit ID, I run git describe to see what upstream version it was released in, so I know what fedora package versions are lacking the fix.

Another tip here is to use the --match option to only search tags matching a particular glob. I've used this to filter out matching against a maintenance or bugfix release branch, when I only wanted to search major version releases.

Don't pull tags from certain remotes

For certain repos like qemu.git, I add a lot of git remotes pointing to individual developer's trees for occasional patch testing. However if trees have lots of non-upstream tags, like for use with pull-requests, they can interfere with my workflow for backporting patches. Use the --no-tags option for this:

 $ git remote add --no-tags $repo