Even though SAGA explicitly does't follow semver conventions, there are
advances, and not all changes on their part break the API. Since the
user is already warned in big letters that versions different from 2.3
are unsupported, in beta status, we've changed the requirement for SAGA
"beta" from 7.3 to everything in version 7.
This change "dims" the results from the SAGA provider when a search
is made in the toolbox, to visually push users towards picking alternative
algorithms instead.
The Processing implementation of SAGA algorithms are a constant source
of critical bugs for users, causing incorrect analysis results. There's
zero community interest in actively maintaining this provider, so we
need to take steps to push users to stop picking these algorithms
wherever alternative (QGIS/GRASS/GDAL based) equivalents exist.
And for 4.0, seriously re-consider dropping this provider from the
out of the box install. We are causing more harm then good by offering
it to users.
Fixes a regression bug accidentally introduced with 40134d6473fcdbd8b6f53c3ea3db01e2dd606419 (PR #8968) and backported with a887b7d34bfa44a8400bcaa986ede96e15a760c9 (PR #9231): when there are multiple output parameters processAlgorithm incorrectly generates multiple saga_cmd commands, instead of a single command containing the output parameters.
Fixes#33658
but show a huge obnoxious "NOT OFFICIALLY SUPPORTED" warning
SAGA 7.3 is the new SAGA LTR, so we eventually need to move towards this.
But adding full support is going to be a painstaking tedious process. So
allow users to use this combination, but totally at their own risk.
TODO: Make SAGA provider a 3rd party plugin only and save QGIS core dev sanity ;)
Looks like there is no problem now in SAGA when using filenames with non-ascii chars, so some code used for replacing filenames has been removed. Also some code for python2, which is not needed now.
fixes#18617
