Closed unhammer closed 1 year ago
ok so it breaks tests :) Is it an actual failure? a tag changed from pl to sg
In the first test that fails, there are 2 rules and they both change the same variable, and I think something might be going wrong with the ordering of those operations.
OK so it seems all the tests run in processGLR
, I thought xmls used processTRX
(can we rename that function to processLinear
or something?)
Think I figured it out (didn't notice that checkForReduce changed currentBranch), tests pass now. @mr-martian does it seem OK to merge?
as noted in https://github.com/apertium/apertium-recursive/issues/80#issuecomment-1691484763 this tries to do a reparse of single, unparsed words after full parse.
It seems to fix things in corpus runs.
Could it break anything @mr-martian ?