case 3977

ayanova/devdocs/todo.txt

@@ -11,24 +11,7 @@
 
 ## ROUGH SCHEDULE TO GET TO BETA
 
-		
+	
-
-   search bug(s):
-	new or possibly always there bug besides mqt thing:
-	search for "007m1qt *nd" matches only two of the three 007m1qt records when it should be three and the *nd matches a bunch of other shit with no 007m1qt in it
-	in v7 there were similar weirdnesses I've noticed when using wildcards mixed with non wildcards or multiple wildcards it can return records without all teh search terms
-	What I expect to happen is records that contain all the search terms only will be returned
-	I'm thinking of rewriting it all as a direct raw sql query as it's tortuous with the complex ef queries required and they don't quite make sense and my tentative semi fix involved just a stab in the dark without fully seeing why it was required
-	then I can troubleshoot and refine entirely in sql and also match more the v7 query cycle which is much simpler
-	maybe all in a stored procedure / view or something?
-	then I could just call the view from ef and it would be ready to roll
-	Make it happen properly.
-	TODO: Each term should narrow down the results to only include those items with *all* the matches in them only regardless of use of wildcards or anything else.
-		prototype it in pgadmin as a single query / view until it works with the combos we need then make it solid and recode that shit
-
-		This query illustrates the problem perfectly: *1QT* *44*
-		which should return exactly one result but returns 12, a bunch of 44's and a few qt's but not AND, rather OR
-
 	Testing cases
 
 	Update front / back 
@@ -782,6 +765,6 @@ BUILD 133 CHANGES OF NOTE
 - case 3996
 - case 3999 see note
 - case 4001
-- case 3977
+- case 3977 fixed