ECCE 8.18.8

Bug fix, backported from 9.0.

• ORCA input, explicit basis sets (#239): a generally contracted shell (Dunning cc-pV*Z, ANO sets) was written with only its first contraction when ECCE wrote the basis set into the ORCA input as explicit primitives. Such calculations lost basis functions without warning and gave wrong energies: water with cc-pVDZ came out 0.31 Eh too high. Calculations where ECCE passed an ORCA basis keyword (for example def2-SVP) were not affected. Affected since ORCA support was added (8.0.7). If you ran ORCA with an explicitly written Dunning or ANO basis, rerun those calculations.

Packages, each built on the system it is for:

File	Built on	Install with
ecce_8.18.8_amd64.deb	Debian 13.7 (trixie)	sudo apt install ./ecce_8.18.8_amd64.deb
ecce_8.18.8_ubuntu24.04_amd64.deb	Ubuntu 24.04	sudo apt install ./ecce_8.18.8_ubuntu24.04_amd64.deb
ecce-8.18.8-1.x86_64.rpm	Rocky Linux 9	sudo dnf install epel-release && sudo dnf install ./ecce-8.18.8-1.x86_64.rpm

Debian 13 tested: the fix is verified with ORCA 6.1.1 (explicit and keyword cc-pVDZ inputs give the same function count and energy); the teaching calculation set passes on this release. The Ubuntu and RPM packages were built on their systems but not tested for this release. RHEL has no ActiveMQ package; see GETTING_STARTED: ActiveMQ on RHEL, Rocky and Fedora.