cppcms1
Build times (auto-updated)
- i686: 2m57s (2026-09-24 08:09 UTC, MAKEFLAGS=-j6)
- pentium4: 3m12s (2026-09-24 08:12 UTC, MAKEFLAGS=-j6)
- x86_64: 3m5s (2026-09-24 08:16 UTC, MAKEFLAGS=-j6)
- New package, created 2026-09-24 as the frozen 1.2.1 legacy fork of
memory/cppcms.md, which now tracks the 2.x line instead. Same build as the
pre-split cppcms (python2-forcing sed hack, -DCMAKE_CXX_STANDARD=17 for
the removed std::allocator::allocate two-arg overload – see cppcms.md for
that history), just renamed/pkgrel reset to 1. conflicts/provides mirror
cppcms so only one of the two can be installed at once (see cppcms.md for
why true side-by-side isn’t possible).
- Git history: rather than starting empty, grafted
cppcms’s full history onto
this repo via a local git fetch ../../../arch/maintained/cppcms master +
git reset --hard FETCH_HEAD (no push from cppcms involved – just copies
existing commit objects), then committed the rename as a new commit on top. So
cppcms1’s AUR log correctly shows the full lineage back through “Initial
import” rather than looking like an unrelated fresh package.
- Same
namcap-flagged missing runtime deps as cppcms 2.0.1 (see that file
for the general finding): icu (booster auto-links it since icu-dev is
present at build time) and python2 specifically here, since this build still
sed-forces the cppcms_tmpl_cc shebang to python2. Both added to
depends=.
- 2026-09-24 build/release: i686, pentium4, x86_64 built+signed+ published
cleanly. i486 fails with
target not found: python2 – expected, same
longstanding chroot-scope gap documented in cppcms.md’s pre-split history:
python2 was only ever built for x86_64 in the archlinuxaba private repo,
not for the 32-bit archs.