And require that showing help is opt-in. Apart from a handful
of built-in providers, most providers will not have help pages
available within the QGIS documentation (including model and
script algorithms). Accordingly, we should hide the help button
by default and only show it for these selected providers.
Note that 3rd party algorithms can still specify custom helpUrl
urls, in which case the button WILL be shown.
Also:
Thanks to @agiudiceandrea:
Fixes#19900
Statistics were erroneously computed on the dissolve field, rather than on the statistic attribute field as it should be.
Fixes#19307
Handle filenames with white spaces (in the same way as Buffer, OffsetCurve, OneSideBuffer, PintsAlongLines algs).
processing options are changed
Make sure the algorithm dialogs use their own copy of algorithms,
instead of the copies owned by the processing registry. Opening
processing options triggers a full reload of providers, deleting
all existing algorithm instances from the registry first.
data defined buttons for model child algorithms
The context was not previously exposed, so users would not have
been aware that they can utilise all the variables and functions
available to parameters within child algorithms.
for pre-calculated expressions
Correctly exposes ALL the variables and functions available for use
in pre-calculated expressions so that users actually know they can
use these in their models!
as well as just variable names
In some cases contexts may provide specific functions of use
to that context, or more generally there may be functions we want
to highlight for a particular expression builder (e.g. highlighting
to_dms in the grid annotation builder)
The context contains settings which reflect the context ini
which a Processing parameter widget is shown, e.g., the
parent model algorithm, a linked map canvas, and other relevant
information which allows the widget to fine-tune its behavior.
the original layer's feature
Otherwise python processFeature implementations may return a
modified shallow copy of this feature, which means that the
test for modifications fails, since we're comparing against
the same feature which was modified.
Fixes some python algorithms do not modify features when used
in-place mode.
