[metzm]

Up to G85, there was a inconsistency in the behaviour of r.reclass.area: mode=lesser method=rmarea removed areas smaller than value, whereas mode=lesser method=reclass removed areas larger than value.

PR #7337 fixed this inconsistend behaviour with the introduction of the new options lower and upper, replacing the mode and value options up to G85. This is great! But handling of backwards compatibility did not take into account the inconsistency in the behaviour up to G85. This PR fixes the backwards compatibility of PR #7337.

[ninsbl]

Thanks for the fix. I overlooked the inconsistency. Should this be mentioned in the manual, also your comment on the rmarea method?

[metzm]

Thanks for the fix. I overlooked the inconsistency. Should this be mentioned in the manual, also your comment on the rmarea method?

IMHO, the previous inconsistency does not need to be mentioned in the manual since you fixed it with the new options. The current behaviour, when are areas removed and when are areas merged with neighbours, should be added to the manual. The name of the module itself, r.reclass.area, does not give a hint about what happens with areas outside the given threshold(s) and needs explanation.

[metzm]

@ninsbl You are welcome to improve my changes to the manual regarding the new lower option.

[ninsbl]

Looks good. Just some autocomplete issues from your IDE...

[echoix]

Also: make sure the typos in the title and the commit message are fixed before merging (relclass -> reclass)

[metzm]

@echoix typos fixed in 0850a40