Merge rc187-cfix: -C without -S keeps the space-group search inside the typed crystal system; Qt notice states the static Linux build

This commit is contained in:
leonarski_f committed 2026-11-15 17:18:44 +01:00
commit a5ee98a285
2 files changed
+16 -6

No files matched your search

+11 -5
View File
@@ -10,12 +10,18 @@ LGPL-3.0 compliance for the distributed viewer:
* Qt is used as provided by The Qt Company under the LGPL-3.0 open-source terms; no Qt
source code has been modified.
* The viewer links Qt dynamically. On Windows and macOS the Qt shared libraries are
deployed next to the application by windeployqt / macdeployqt; the user may therefore
replace them with a compatible build of Qt, as the LGPL requires.
* On Windows and macOS the viewer links Qt dynamically: the Qt shared libraries are
deployed next to the application by windeployqt / macdeployqt, and the user may replace
them with a compatible build of Qt, as the LGPL requires.
* The Linux build links Qt 6.11.x (currently 6.11.1) statically. As LGPL-3.0 section 4(d)(0)
provides, the user can relink the viewer with a modified Qt: the complete source of
jfjoch_viewer is distributed with Jungfraujoch, and docker/ubuntu2404/Dockerfile (with
docker/common/build-qt.sh) gives the full recipe that builds the static Qt and the viewer;
point it at the modified Qt and rebuild jfjoch_viewer.
* The corresponding source code for the exact version of Qt used can be obtained from
https://download.qt.io/ (and from https://code.qt.io/). The Qt version is reported in
the viewer's Help > About dialog.
https://download.qt.io/ (the Linux build fetches
https://download.qt.io/official_releases/qt/6.11/6.11.1/single/qt-everywhere-src-6.11.1.tar.xz)
and from https://code.qt.io/. The Qt version is reported in the viewer's Help > About dialog.
Full license texts:
* LGPL-3.0: see licenses/Qt6-LGPL-3.0.txt (verbatim GNU Lesser General Public License v3),
+5 -1
View File
@@ -2836,7 +2836,11 @@ bool Rugnux::ScaleMergeAndSymmetry(PipelineLocals &p) {
// answered by its having run. The two cannot disagree: the ask below only ever replaces the
// answer in hand with a STRICTLY higher point group, and it writes all four commit_ values
// together, so the group, the cell and the lattice type stay one answer.
if (end_msg.rotation_lattice.has_value() && end_msg.rotation_lattice_type.has_value()) {
//
// A cell given with -C fixes the crystal system, and the indexer already holds it there, so
// a higher class is never asked: the search stays within the system the cell names.
if (end_msg.rotation_lattice.has_value() && end_msg.rotation_lattice_type.has_value()
&& !experiment_.GetIndexingSettings().GetFixedLatticeSystem()) {
TimingMark metric_mark(logger, "metric lattice re-ask");
// The merge is indexed on the lattice the indexer named, whatever has been committed
// since, so the metric is measured on the primitive form of THAT one.