Order of tiled plots

Expected behavior

If I do a tiled plot of a workspace group, the subplots should be in the same order as the workspaces in the group.

Actual behavior

The tiles are in alphabetical order of child workspace name. For the output from a simultaneous fit, where there are more than 10 spectra, this is FitName_Workspace_0, 1, 10, 11, …, 19, 2, 20, 21, … 9. This is somewhat confusing, especially if the workspaces were supposed to be ordered in increasing temperature, magnetic field or similar.

Steps to reproduce the behaviour

Do a simultaneous fit with 11 or more spectra (or otherwise create a group where the names are not strictly alphabetical, and workspaces have 2 or more spectra each). Right click - Plot spectra - Tiled plot - Plot All. Compare the tile order with that in the Workspaces pane when the group is expanded.

Platforms affected

Windows (plus probably others), noticed in 3.9.2.

As an aside - it isn’t possible to get a tiled plot this way if the child workspaces only have one spectrum each. The only option is to have them all superimposed. Often that isn’t very useful. It is possible to get a tiled plot with one plot per tile, for example spectrum 0 from each workspace, if there are more spectra.

To follow up: in Muon Analysis, a simultaneous fit across multiple runs produces a workspace group where the workspaces are named with the run number and so do display in the right order. But across detector groups, they are sorted by group name rather than position in the grouping table. But that sorting takes place earlier in the interface since even the dialog box to enter individual parameters shows the groups in sorted order.