ICU-21179 Clean up unused code in the RBBI dictbe

The wordsMatched conditional gave the misleading impression that only the first "first word" candidate that was followed by a dictionary word would be marked, but because the wordsMatched declaration is inside the while loop, every "first word" candidate that's followed by a dictionary word is marked (resulting in the shortest and not the longest first candidate that's followed by a second dictionary word).
diff --git a/icu4c/source/common/dictbe.cpp b/icu4c/source/common/dictbe.cpp
index b42cdf0..4428575 100644
--- a/icu4c/source/common/dictbe.cpp
+++ b/icu4c/source/common/dictbe.cpp
@@ -265,13 +265,9 @@
                 goto foundBest;
             }
             do {
-                int32_t wordsMatched = 1;
                 if (words[(wordsFound + 1) % THAI_LOOKAHEAD].candidates(text, fDictionary, rangeEnd) > 0) {
-                    if (wordsMatched < 2) {
-                        // Followed by another dictionary word; mark first word as a good candidate
-                        words[wordsFound%THAI_LOOKAHEAD].markCurrent();
-                        wordsMatched = 2;
-                    }
+                    // Followed by another dictionary word; mark first word as a good candidate
+                    words[wordsFound%THAI_LOOKAHEAD].markCurrent();
                     
                     // If we're already at the end of the range, we're done
                     if ((int32_t)utext_getNativeIndex(text) >= rangeEnd) {
@@ -503,13 +499,9 @@
                 goto foundBest;
             }
             do {
-                int32_t wordsMatched = 1;
                 if (words[(wordsFound + 1) % LAO_LOOKAHEAD].candidates(text, fDictionary, rangeEnd) > 0) {
-                    if (wordsMatched < 2) {
-                        // Followed by another dictionary word; mark first word as a good candidate
-                        words[wordsFound%LAO_LOOKAHEAD].markCurrent();
-                        wordsMatched = 2;
-                    }
+                    // Followed by another dictionary word; mark first word as a good candidate
+                    words[wordsFound%LAO_LOOKAHEAD].markCurrent();
                     
                     // If we're already at the end of the range, we're done
                     if ((int32_t)utext_getNativeIndex(text) >= rangeEnd) {
@@ -699,13 +691,9 @@
                 goto foundBest;
             }
             do {
-                int32_t wordsMatched = 1;
                 if (words[(wordsFound + 1) % BURMESE_LOOKAHEAD].candidates(text, fDictionary, rangeEnd) > 0) {
-                    if (wordsMatched < 2) {
-                        // Followed by another dictionary word; mark first word as a good candidate
-                        words[wordsFound%BURMESE_LOOKAHEAD].markCurrent();
-                        wordsMatched = 2;
-                    }
+                    // Followed by another dictionary word; mark first word as a good candidate
+                    words[wordsFound%BURMESE_LOOKAHEAD].markCurrent();
                     
                     // If we're already at the end of the range, we're done
                     if ((int32_t)utext_getNativeIndex(text) >= rangeEnd) {
@@ -908,13 +896,9 @@
                 goto foundBest;
             }
             do {
-                int32_t wordsMatched = 1;
                 if (words[(wordsFound + 1) % KHMER_LOOKAHEAD].candidates(text, fDictionary, rangeEnd) > 0) {
-                    if (wordsMatched < 2) {
-                        // Followed by another dictionary word; mark first word as a good candidate
-                        words[wordsFound % KHMER_LOOKAHEAD].markCurrent();
-                        wordsMatched = 2;
-                    }
+                    // Followed by another dictionary word; mark first word as a good candidate
+                    words[wordsFound % KHMER_LOOKAHEAD].markCurrent();
 
                     // If we're already at the end of the range, we're done
                     if ((int32_t)utext_getNativeIndex(text) >= rangeEnd) {