last month's status: https://firstname.lastname@example.org/thread/6...
Last months' reproducible builds project updates (including my work): https://reproducible-builds.org/reports/2021-03/
I uploaded https://rb.zq1.de/compare.factory-20210428/ today and rbstats are: total-packages: 13955 (+87) build-tried: 13945 (+86) build-failed: 63 (+28) build-n-a: 173 (+20) build-succeeded: 13709 (+38) build-official-failed+na: 262 (-80) build-compare-failed: 366 (-57) build-compare-succeeded: 13343 (+95) verify-failed: 495 (+2) verified-semi-reproducible: 12960 (+72) bit-by-bit-identical: 13191 (+72) not-bit-by-bit-identical: 535 (-17) not-bit-by-bit-identicalcheck: 518 (-34)
https://rb.zq1.de/compare.factory-20210428/graph.png shows the change over time
https://rb.zq1.de/compare.factory-20210428/unreproduciblerings.txt lists very unreproducible core packages (bootstrap+DVD)
Of the badly unreproducible packages, 3 were in ring0 45 were in ring1
That makes it 48/3247 => 1.48 % which is below the overall average of 366/13709 => 2.67 %
535/13709 => 3.90 % of packages are not perfectly reproducible
This month I only did some easy patches as I have been more busy with reproducibility of 15.3 and sysadmin work at SUSE.
On the list of unreproducible DVD packages, fltk is unreproducible again from doxygen ASLR - maybe a regression on https://github.com/doxygen/doxygen/issues/7474
qemu is there again with some ordering issues in its binaries
and there are still dozens of packages suffering from xmvn https://bugzilla.opensuse.org/show_bug.cgi?id=1162112 and mono toolchain issues https://bugzilla.opensuse.org/show_bug.cgi?id=1141502 Would be nice if someone could help there.
In general, I still see a lot of python. Not only because .pyc files are not designed for determinism, but also because python is one of the few languages where glob("*") is non-deterministic. And people use os.listdir and os.walk too.
At least there was some progress in upstream java with https://github.com/openjdk/jfx/pull/446 though it is not merged yet.
Ciao Bernhard M.