This commit provides a different help landing page (Settings) for preferences that are used outside the normal Ctrl+P preferences dialog - for example, 'spectrum settings' accessed from the track menu. This should work for any settings that are used on their own in that way, though spectrum settings is currently the only example.
The wave is now green, to match the known-good black/green contrast for this theme. Corresponding changes were made to the buttons, slider pips, label shading and draggers. Time before zero is now also a lighter blue which shows up better against black than the dark blue did. Sliders are now (almost) white rather than grey. Selection in label tracks now matches selection in wave tracks. Envelopes are now (almost) white.
Changed the criterion for deciding when two clips are immediately next to each other, and made it into a function: WaveClip::SharesBoundaryWithNextClip.
That is, don't assume uniformity of pixels per second, but supply the correct
first argument.
This will matter when there is a magnifier. But that will not be in the next
version.
This means you can now select just a time, and get all tracks selected, or some tracks, and get all time selected.
One side effect of this is that if you select a label (only) and apply 'cut' you now just cut the label, rather than cutting all tracks. If you select a label and apply echo, you now get an error message, rather than applying echo to all tracks.
... also gathered various formatting into some routines. Should we regularize
the messages further? That is, consistently put "dB" before or after "RMS"
Also added more i18n-hint comments
Now if there is a time selection and no tracks selected, then just select all the tracks, preserving the time selection.
This helps in the case a user has made a time selection, e.g. with selection toolbar, and then clicked on track panel, losing the selection of tracks but preserving the time selection.
I also shortened some repeated cut-and-pasted code.
Fix bug in AudacityProject::AdjustForFindingStartTimes.
Bug: When there is 0 or 1 clips, q - 2 is before the begin() of the vector, and caused crash.
Fix: split the test into two tests.
