CI currently failing due to #7179

This indeed feels like band-aid fixing, but better than what we have, so why not. Perhaps a TODO/FIXME comment in the code itself would be welcome. msSLDGetLogicalOperator() would also benefit from a similar logic.
Ideally we'd perhaps need to use yylex() and a new SLD parsing mode in mapparser.y

This indeed feels like band-aid fixing, but better than what we have, so why not. Perhaps a TODO/FIXME comment in the code itself would be welcome. msSLDGetLogicalOperator() would also benefit from a similar logic. Ideally we'd perhaps need to use yylex() and a new SLD parsing mode in mapparser.y

Yes, these changes are very much trying to make an error-prone approach slightly less buggy. With further tests it appears both msSLDGetLogicalOperator, msSLDGetRightExpressionOfOperator, and msSLDGetLeftExpressionOfOperator have the same issues - see the test where "An And value" is converted to "An".

@sdlime - could any of the changes in https://mapserver.org/development/rfc/ms-rfc-64.html be used here instead of the current string searching approach?

I've a copy of the Lex Yacc book but never got much past the first chapter..

I've updated the function so the previous cases such as AND( and OR( still pass, so anything that previously worked should continue to do so. I keep thinking that expression parsing is implemented somewhere - msEvalTextExpressionInternal seems close but returns a true/false rather than a parse tree that could be used for converting expressions from Mapfile to SLD.