where did figleaf_htmlizer.py come from?

[zooko]

I removed src/allmydata/util/figleaf_htmlizer.py@3615 from the Ubuntu Karmic package so that I wouldn't have to explain its licensing terms and provenance. I also removed it from Tahoe-LAFS trunk so that Tahoe-LAFS trunk would be closer to the Ubuntu Karmic package. aaaa633f189adba4 This broke our figleaf build, as I expected it would: ​http://allmydata.org/buildbot/builders/edgy/builds/2302/steps/figleaf-html/logs/stdio

I figured I would then update our Makefile to invoke the upstream figleaf library from ​http://github.com/ctb/figleaf/blob/master/doc/index.txt . However, when I tried to do so I discovered that the figleaf_htmlizer.py doesn't seem to have an analog in the upstream figleaf tarball, and in particular figleaf upstream doesn't seem to offer the -o functionality that we rely on to compute the delta in coverage.

So, where did figleaf_htmlizer.py? Was it authored by Brian? In that case we can legally include it in Tahoe-LAFS under the same licensing terms as the rest of Tahoe-LAFS, without having to explain ourselves to Ubuntu. The revision control history starts with: e8ee3365eff0dbb1, which copies it into our repo from an unspecified source.

Perhaps we could persuade the upstream figleaf project to accept it into their repo. In any case, our figleaf builder is currently broken, and before we put figleaf_htmlizer.py back into our repo to fix it, I would like to know if there is some third-party source for it so that we don't have a licensing issue with Ubuntu (and ideally so that we can share maintenance of it with other projects).

[zooko]

I opened a ticket on the figleaf project asking them to adopt this module:

​http://github.com/ctb/figleaf/issues/#issue/1

[zooko]

Okay the figleaf project hasn't replied to the ticket I opened. Brian: did you write figleaf_htmlizer.py? If so, great, we can put it back. Let's just put something at the top like:

written by Brian Warner
copyright 2009 allmydata.com
This file is part of the Tahoe-LAFS project and can be used under the same terms as the project, see docs/about.html for details

Then we can close this ticket and our figleaf service will work again, and we can open a new ticket saying something like "Persuade the upstream figleaf project to adopt our changes".

Assigning this to Brian because only he can state that he wrote it and that we can have it under the Tahoe-LAFS licences, but I'm willing to do the legwork of putting it back and making the figleaf builders work again.

[zooko]

I wrote to figleaf maintainer Titus Brown and he said he hadn't seen ​http://github.com/ctb/figleaf/issues/#issue/1 but he would look into it.

[zooko]

If you like this ticket, you may also like #109 (current upstream figleaf doesn't do the right thing), #387 (figleaf confused by python2.5), and #623 (figleaf: compute+display differential coverage).

[zooko]

Brian: just say "I wrote figleaf_htmlizer.py and Tahoe-LAFS can license it however Zooko wants". Or say "I derived it from upstream figleaf project." or whatever: just bring to an end the mystery of its provenance and licensing please!

[zooko]

I no longer care where figleaf_htmlizer.py came from because #930 (switch from figleaf to coverage.py).