last month's status:
Last months' reproducible builds project updates (including my work):
I made some nice progress with Firefox+Thunderbird (see below).
This month also saw the enabling of Link Time Optimization (LTO) in
Factory and this triggered a class of unreproducibility because the
number of build CPUs was added to the CFLAGS value.
As a quick workaround I added
so that it does not reduce visibility of other reproducibility issues.
are related. A nicer solution still needs to be found.
Nearly fixed core packages:
contains a working patch
newly unreproducible core packages:
similar to https://bugzilla.gnome.org/show_bug.cgi?id=794069
maybe was not really fixed last month?
half-building packages that disturb my tools:
I uploaded https://rb.zq1.de/compare.factory-20190726/
and rbstats (with comparison to last month) are:
total-packages: 12096 (+82)
build-tried: 12086 (+82)
build-failed: 48 (+0)
build-n-a: 119 (+13)
build-succeeded: 11919 (+69)
build-official-failed+na: 208 (+35)
build-compare-failed: 352 (+4)
build-compare-succeeded: 11567 (+65)
verify-failed: 560 (+29)
verified-semi-reproducible: 11279 (+33)
verified-bit-identical: 0 (+0)
bit-by-bit-identical: 11257 (+68)
not-bit-by-bit-identical: 659 (+8)
shows the change over time
lists very unreproducible core packages (bootstrap+DVD)
Of the badly unreproducible packages,
3 were in ring0
40 were in ring1
That makes it 43/2901 => 1.48 %
which is below the overall average of
352/11919 => 2.95 %
659/11919 => 5.53 % of packages are not perfectly reproducible