IRC channel logs
2019-09-23.log
back to list of logs
<janneke>vagrantc: no, XPASS is an uneXpected pass; the test is listed in FAIL_TESTS, but passes <janneke>vagrantc: yes, mes touches to epoch for reproducibility <janneke>that's weird, i wonder what #debian-reproducible thinks about that :) <vagrantc>in Debian, packages are built with SOURCE_DATE_EPOCH set to the last debian/changelog version, and maybe even clamps files from the source newer than that. <vagrantc>but guix's habit of making all files have the epoch set to 0 is a bit different approach <janneke>vagrantc: ah yes, any fixed date will do. it's just less work for me to reproducibly calculate date==0 :-) <vagrantc>janneke: Debian also assumes reproducibility of the source tarball as an input where the dates will be set in the tarball already... <vagrantc>janneke: i might have to produce a mangled tarball with the files marked to the date of commit or something. <vagrantc>janneke: e.g. you pass the same tarball every time, so the dates of all the files within the tarball aren't going to vary. <janneke>vagrantc: would you like to align that strategy with (thhe debian folks@) #reproducible-builds on oftc? possibly a naive question may help. <vagrantc>janneke: probably better to post to one of the mailing lists <vagrantc>janneke: are you subscribed to rb-general? <vagrantc>janneke: ok, i'll start the timestamps in source tarballs discussion as relates to debian packaging there