What happened in the Reproducible Builds effort between June 26th and July 2nd 2016:

Read on to find out why we're lagging some weeks behind…!

GSoC and Outreachy updates

  • Ceridwen described using autopkgtest code to communicate with containers and how to test the container handling.

  • reprotest 0.1 has been accepted into Debian unstable, and any user reports, bug reports, feature requests, etc. would be appreciated. This is still an alpha release, and nothing is set in stone.

Toolchain fixes

  • Matthias Klose uploaded doxygen/1.8.11-3 to Debian unstable (closing #792201) with the upstream patch improving SOURCE_DATE_EPOCH support by using UTC as timezone when parsing the value. This was the last patch we were carrying in our repository, thus this upload obsoletes the version in our experimental repository.
  • cmake/3.5.2-2 was uploaded by Felix Geyer, which sorts file lists obtained with file(GLOB).
  • Dmitry Shachnev uploaded sphinx/1.4.4-2, which fixes a timezone related issue when SOURCE_DATE_EPOCH is set.

With the doxygen upload we are now down to only 2 modified packages in our repository: dpkg and rdfind.

Weekly reports delay and the future of statistics

To catch up with our backlog of weekly reports we have decided to skip some of the statistics for this week. We might publish them in a future report, or we might switch to a format where we summarize them more (and which we can create (even) more automatically), we'll see.

We are doing these weekly statistics because we believe it's appropriate and useful to credit people's work and make it more visible. What do you think? We would love to hear your thoughts on this matter! Do you read these statistics? Somewhat?

Actually, thanks to the power of notmuch, Holger came up with what you can see below, so what's missing for this week are the uploads fixing irreprodubilities. Which we really would like to show for the reasons stated above and because we really really need these uploads to happen ;-)

But then we also like to confirm the bugs are really gone, which (atm) requires manual checking, and to look for the words "reproducible" and "deterministic" (and spelling variations) in debian/changelogs of all uploads, to spot reproducible work not tracked via the BTS.

And we still need to catch up on the backlog of weekly reports.

Bugs submitted with reproducible usertags

It seems DebCamp in Cape Town was hugely successful and made some people get a lot of work done:

61 bugs have been filed with reproducible builds usertags and 60 of them had patches:

Package reviews

437 new reviews have been added (though most of them were just linking the bug, "only" 56 new issues in packages were found), an unknown number has been been updated and 60 have been removed in this week, adding to our knowledge about identified issues.

4 new issue types have been found:

Weekly QA work

98 FTBFS bugs have been reported by Chris Lamb and Santiago Vila.

diffoscope development

strip-nondeterminism development

  • Chris Lamb made sure that .zhfst files are treated as ZIP files.

tests.reproducible-builds.org

  • Mattia Rizzolo uploaded pbuilder/0.225.1~bpo8+1 to jessie-backports and it has been installed on all build nodes. As a consequence all armhf and i386 builds will be done with eatmydata; this will hopefully cut down the build time by a noticeable factor.

Misc.

This week's edition was written by Mattia Rizzolo, Reiner Herrmann, Ceridwen and Holger Levsen and reviewed by a bunch of Reproducible builds folks on IRC.