Will's blog

purpose: Will Kahn-Greene's blog of Python, Linux, random content, PyBlosxom, Miro, and other projects mixed in there ad hoc, half-baked, and with a twist of lemon

[ home | blog home | recent activity | guestbook | plugins i'm using (19) ]

Page 1 of 73  >> (less recent)

Wed, 01 Sep 2010

Dev call 9/1/2010 minutes

Miro 3.5 status (roadmap)

Miro Community 1.1.1 status (roadmap)

Miro Community 1.2 status (roadmap)

Paul:

Janet:

Luc:

Ben:

Will:

Order of business

Bugzilla stats for Miro:

Fri, 27 Aug 2010

Server shuffling

We've had a few servers are a hosting facility in Marlboro, MA, USA for a long time now. Over the last year, we've been moving applications, sites, systems and all that from the servers in Marlboro to the Amazon cloud system. We're in the process of moving the last set of things to Amazon.

We're working on making the transition go smoothly, but there are a lot of pieces and it's possible we'll miss something along the way. If you notice PCF/Miro-related sites being down, let me know.

Wed, 25 Aug 2010

Dev call 8/25/2010 minutes

Miro 3.5 status (roadmap)

  • working on fixing up video conversion issues
  • working on cleaning things up

Miro Community 1.1.1 status (roadmap)

  • Miro Community 1.1 was released last week
  • working on some bug fixes targeted for 1.1.1

Paul:

  • pushed Miro Community 1.1 out to the new servers
  • looking at moving the db to another machine

Ben:

  • continued working on bugs for Miro 3.5
  • implemented font selection for subtitles on Windows

Will:

  • implemented the PCF bug triage Firefox addon
  • worked on database startup failure problems

Janet:

  • working on Windows installer things
  • and other things

Order of business

  • We talked a little about what Miro 4 will entail. We're going to try to plan it out in a wiki rather than plan it out in Bugzilla because Bugzilla is difficult to use for iterative design. When tasks become sufficiently defined, we'll move them to Bugzilla and start implementing.
  • Did you know there's a Miro User Manual? If you haven't looked at it yet, it's worth taking a look at. You can find it at http://getmiro.com/userguide/.
  • Priority is being put on finishing development work on Miro 3.5. Any help with this would be huge. There are a series of bugs on the roadmap that need to be fixed. If you want to help out but don't know where to start, see http://getmiro.com/userguide/contributing.html
  • If you can't contribute work, please consider contributing funding by donating. Your money goes directly to ongoing development of Miro and related projects.

Bugzilla stats for Miro:

Lot's of bug churn this week in part because of focus on regression testing and because of the creation (and heavy usage of) the PCF bug triage Firefox addon.

  • 20 bugs/feature-requests created
  • 3 bugs marked WORKSFORME
  • 1 bugs marked INVALID
  • 3 bugs marked DUPLICATE
  • 1 bugs marked WONTFIX
  • 12 bugs marked FIXED
  • 30 bugs marked INCOMPLETE

Wed, 18 Aug 2010

Dev call 8/18/2010 minutes

Miro 3.5 status (roadmap)

  • working on fixing up video conversion issues
  • working on cleaning things up

Miro Community 1.1 status (roadmap)

  • working on it--probably a release today or tomorrow

Paul:

  • Planning to tag Miro Community 1.1 and pushing out the release in the next day or two. Yay!
  • Worked on the Miro installer for Windows.

Ben:

  • Worked on bugs for Miro 3.5.
  • Talked about how to tackle bug 14203.

Will:

  • Worked on bugs.
  • Tested Miro on Maverick.

Order of business

  • We spent some time talking about fundraising. We're trying to get to the point where Miro development is self-funding.
  • Did you know there's a Miro User Manual? If you haven't looked at it yet, it's worth taking a look at. You can find it at http://getmiro.com/userguide/.
  • Priority is being put on finishing development work on Miro 3.5. Any help with this would be huge. There are a series of bugs on the roadmap that need to be fixed. If you want to help out but don't know where to start, see http://getmiro.com/userguide/contributing.html
  • If you can't contribute work, please consider contributing funding by donating. Your money goes directly to ongoing development of Miro and related projects.

Bugzilla stats for Miro:

  • 13 bugs/feature-requests created
  • 1 bugs marked WONTFIX
  • 7 bugs marked FIXED
  • 1 bugs marked WORKSFORME
  • 1 bugs marked DUPLICATE
  • 1 bugs marked INCOMPLETE

Fri, 13 Aug 2010

New Windows build environment

I spent a good amount of time over the last few months migrating Miro on Windows to a new Windows build environment that uses Python 2.6.5 and Visual C++ 9.0 (part of Visual Studio 2008).

I landed the changes two weeks ago. Janet, Ben and I identified a couple of problems and sorted those out. Last week I got the Windows build box to produce nightlies without requiring any babysitting.

Yay!

Features of the new Windows build environment:

  1. It's a lot easier to configure. Previously, we had to set PATH, INCLUDE, and LIB environment variables to the right magic values to build correctly. Python 2.6 automatically pulls those values from Visual Studio 2008 files. So we don't need to deal with those anymore.
  2. It uses Visual C++ 9.0 (Visual Studio 2008) rather than VC++ 7.1 (VS 2003). The latter can no longer be acquired legally (thank you MS). There's an Express version of the former available on the Microsoft site for free.
  3. We were using Python 2.5 which had a bunch of bugs we had workarounds for. We're now using Python 2.6.5 which doesn't have these issues and also has a series of optimizations that should make Miro run better.
  4. There's a get_requirements.sh script that downloads the versions of Python and libraries that you need automatically.

This also means that we can require Python 2.6 or later on all platforms for Miro. Therefore we can:

The one thing we still want to do is upgrade from gtk 2.16 to gtk 2.20. Bug 14037 covers the problems here. We're blocked by Bug 625972 in gtk.

I've written up instructions on setting up the new Windows build environment. It takes me about 30 minutes to do--mostly because it takes a while to install Visual C++ 9.0 Express. It's much easier to set up the new environment than the previous environment. When I first started at PCF a few years ago, it took me a couple of days to get the Windows build environment working.

For more details on the new Windows build environment, see the wiki page on WindowsBuildDocs.

Comments have been broken

Rick kindly pointed out that comments on my blog were broken. Totally my fault--I've been working on PyBlosxom on and off and a couple of weeks ago I updated my blog to use the latest version of PyBlosxom in git master and didn't grab the latest comments plugin.

Comments were getting into the moderator queue (I moderate all comments), but when the comments plugin went to send me a notification email, it'd die.

They're working again.

Thu, 12 Aug 2010

PyOhio 2010 videos posted to Python Miro Community

Last night, I spent a few hours going through all the videos from PyOhio 2010, added metadata from the PyOhio site, added notes about video issues, and approved all of the videos except one which I'm talking with Carl about.

It looks like PyOhio 2010 was a really great conference: the talks are really great, you can feel the energy in the rooms from the people who attended, the questions are interesting, and the videos from the conference are fantastic. Also, I'm a huge +1 on using the headset mics--they absolutely make a difference in being able to hear the presenter throughout the video.

Here are three sessions I watched all the way through:

  • Teach Me Python Bugfixing - http://python.mirocommunity.org/video/1794/pyohio-2010-teach-me-python-bu
    This session walks new people through how to contribute to Python core. It's worth watching for new contributors. I also think this is useful for illustrating the barriers to entry and required workflow knowledge for new contributors.
  • PyPy and Unladen-Swallow: Making your Python Fast - http://python.mirocommunity.org/video/1804/pyohio-2010-pypy-and-unladen-s
    Alex did a test-run of this talk at ChiPy. I studied programming languages and compilers in grad school and find these talks really interesting. I liked the analysis of "slow" in respects to the language and the cpython implementation--it's towards the beginning.
  • Wrangling the bits, standardizing how apps get built - http://python.mirocommunity.org/video/1789/pyohio-2010-wrangling-the-bits
    This covers things in virtualenv I didn't know where there. It also talks about modern-package-template which I didn't know existed and fabric which I haven't played with before. Rick also mentions searching Google Code for "after_install" and "fabfile" for virtualenv and fabric usage examples. He says there's a wealth of code snippets in there. I watched the session and immediately started adjusting my scripts. This session had mic problems and the session starts at 7:20. I'm working with Carl to get the first 7:20 clipped.

Sessions from PyOhio 2010 are in the PyOhio 2010 category. Take some time today to browse the sessions and watch the ones that interest you.

Also, I'm sitting on videos for two other conferences: DjangoConf EU and Kiwi PyCon. I'm hoping to get to them in the next couple of weeks. I did PyOhio 2010 videos last night because Carl needed someone to spot-check them to make sure the uploads were good.

Wed, 11 Aug 2010

Dev call 8/11/2010 minutes

Miro 3.5 status (roadmap)

  • working on fixing up video conversion issues
  • working on fixing libcurl related issues

Miro Community 1.1 status (roadmap)

  • working on it

Janet:

  • worked on regression testing for Miro 3.5 on Windows
  • worked on Miro 3.5 bug fix verification
  • worked on Universal Subtitles testing

Ben:

  • worked on Miro 3.5 bugs focusing on downloader-related bugs

Will:

  • fixed Windows build box
  • fixed libtorrent issue on Windows
  • synced translations with Launchpad
  • fixed Miro 3.5 bugs and did some bug triage
  • helped thrillERboy land his first FLOSS patch

Paul:

  • worked on Miro Community issues and development

Order of business

  • Priority is being put on finishing development work on Miro 3.5. Any help with this would be huge. There are a series of bugs on the roadmap that need to be fixed. If you want to help out but don't know where to start, see http://getmiro.com/userguide/contributing.html
  • If you can't contribute work, please consider contributing funding by donating. Your money goes directly to ongoing development of Miro and related projects.

Bugzilla stats for Miro:

  • 14 bugs/feature-requests created
  • 2 bugs marked DUPLICATE
  • 10 bugs marked FIXED

Mon, 09 Aug 2010

Working on bugs: hurdles to bug squashing

There are two big hurdles to get through when starting on a bug before I'm in a position where I can fix it. I think the bulk of the time spent on most bugs is in getting through these two hurdles. If we could somehow reduce the amount of time and energy I spend on getting through the two hurdles, we can increase the speed that I can fix bugs.

Copyright 2008 Flickr user foxtongue

The first hurdle: what is the problem?

I think most people confuse the question "what is the problem" with "what's causing the problem". These are two distinct questions. The first one can only be answered by whoever reported the problem and this hurdle is a communication hurdle. I spend a lot of time trying to coach people into explaining the problem they have in a way that's coherent and complete.

It's not uncommon to get bug reports along the lines of "something's wrong and Miro doesn't work". I'm psyched someone has taken the time to report the issue, but there isn't enough detail here for me to do anything. In order for me to do something with this bug, I spend time walking the reporter through giving me enough details such that I have a good idea as to what the problem is. Because this requires communicating back and forth, this can take anywhere from a few hours to a few weeks to transpire.

Because it takes so long and requires me to ask 20 questions, I'm pretty sure most bug reporters find this phase very frustrating. Some question and answer volleys are really frustrating for me, too. I want to help--I just don't know what the problem is.

Assuming we get past that hurdle, then the second hurdle comes into play....

The second hurdle: what is the cause of the problem?

This hurdle is sometimes difficult because I can't reproduce the problem. I may not have the right context (e.g. "Problem XYZ happens to all people in Ecuador") or the right equipment (e.g. "Problem XYZ happens to people using Windows XP SP2 with video card Foo"). I'll read through code and work with the reporter to try to figure out the cause. Sometimes it works and sometimes it doesn't. When it doesn't it's really frustrating.

Most bugs aren't like that, though. I'm able to either reproduce the context and equipment or I'm able to use a reasonable facsimile. Then figuring out the cause is entirely on my shoulders and I can work through it like I do most things: code spelunking, research, Google searches, talking with co-workers, talking with upstream developers, ....

How can you help?

The following helps a lot:

  • use an email address that you respond to
  • be as specific as you can in your bug report description
  • attach your Miro log file
  • be patient with me when I ask my 20 questions to figure out what the problem is
  • be prompt with responses--the longer it takes to turn around a question and answer, the longer it takes to get the bug fixed
  • thank people who help get your bug fixed: other people who comment on the bug providing additional details, people who worked on the bug, people who tested the bug fix--it takes a community to build an application

Guess what....

This isn't specific to me or Miro! This is true of all FLOSS applications that you use: we're all working hard to build the best applications ever. Doing the above helps everyone.

Wed, 04 Aug 2010

Dev call 8/4/2010 minutes

Miro 3.5 status (roadmap)

  • working on fixing up video conversion issues
  • working on fixing libcurl related issues

Miro Community 1.1 status (roadmap)

  • working on it
  • finishing up implementing playlist features

Janet:

  • having problems with downloading using Windows build
  • worked on Universal Subtitles test setup and testing

Ben:

  • worked on bug 14037. created a test case that fails with gtk 2.20 and submitted the bug 625972 upstream
  • working on creating an OSX build environment

Will:

  • worked on Windows build environment issues
  • downgraded us to gtk 2.16 which doesn't have client-side windows
  • cleaned up and rebuilt the Windows build box to use the new environment
  • started working on better community organization
  • fixed http://planet.getmiro.com/
  • upgraded the Miro User Manual to 3.0.3

Luc:

  • working on video conversion problems in bug 13821
  • worked on other complex video conversion problems

Paul:

  • worked on playlists for MC 1.1
  • worked on the Windows installer

Order of business

  • The Windows build environment work is now done and we're using Python 2.6.5 and VC++ 9.0 (yay!).
  • Priority is being put on finishing off this release. Any help with this would be huge. There are a series of bugs on the roadmap that need to be fixed. If you want to help out but don't know where to start, see http://getmiro.com/userguide/contributing.html
  • If you can't contribute work, please consider contributing funding by donating. Your money goes directly to ongoing development of Miro and related projects.

Bugzilla stats for Miro:

  • 10 bugs/feature-requests created
  • 2 bugs marked FIXED
  • 5 bugs marked WONTFIX
  • 1 bugs marked WORKSFORME
  • 5 bugs marked DUPLICATE
  • 4 bugs marked INCOMPLETE

Page 1 of 73  >> (less recent)


pyblosxom::1.5rc2 20100803

All contents Copyright 1996 to 2010 Will Guaraldi Kahn-Greene.

Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-Share Alike 3.0 United States License.