Sunday, 30 January 2011

Bug reporting energy depleted

Today's lesson: If you search the Python bug tracker with a word directly relevant to a bug you are considering reporting, that you do not get any matches does not mean there isn't a bug squirreled away somewhere.

I'm doing the Stackless merge for 3.2 and one of the problems is a crash in Lib\test\ where Windows pops up a dialog resulting in the pausing of the unit test execution until the user deals with it. I googled for reports of the problem, and I searched the Python bug tracker.

So I proceeded to register an account and submit a bug and seconds later, it was closed as duplicate. All that effort wasted, but considering that my main concern was that the bug would simply sit there and never be looked at, it's a much preferable result.

I went back to the search page, and did a search there on "all text" (which looks at message bodies and titles). No dice.

One last try, the direct title search.

There we go. Somethings rotten in the state of Denmark and all that. I admire the speed at which bugs are dealt with on the Python bug tracker, and I don't know how many bug reports you get let alone duplicates that have to be closed, but you might be able to save yourself some work if you can get the main search gadget doing better searching. I'd submit this to the bug tracker bug tracker, but this has drained too much energy from me which I need to get the Stackless merge out of the way. Passive reporting via Planet Python is the way it is going to have to be.


  1. And done:

  2. Thanks Nick, I was feeling curmudgeonly about not reporting it, but just don't have the time and energy to space.

  3. No worries, I can certainly relate to that feeling - I run into a lot with Kubuntu problems where I can't readily characterise the issue well enough to turn it into a useful bug report.