[pulse] avoid false positives involving `folly::fbstring`

Summary:
That class does some complicated accounting of memory that depends on
whether the string is "small", "medium" or "large". In the latter case
it does its own ref-counting and copy-on-write to save memory, and that
trips up pulse. Pretending all strings are small avoids that issue.

Reviewed By: skcho

Differential Revision: D18909030

fbshipit-source-id: 1c14d909b

infer/src/pulse/PulseModels.ml

@@ -348,6 +348,9 @@ module ProcNameDispatcher = struct
       ; +match_builtin BuiltinDecl.__cast <>$ capt_arg_payload $+...$--> Misc.id_first_arg
       ; +match_builtin BuiltinDecl.abort <>--> Misc.early_exit
       ; +match_builtin BuiltinDecl.exit <>--> Misc.early_exit
+      ; (* consider that all fbstrings are small strings to avoid false positives due to manual
+           ref-counting *)
+        -"folly" &:: "fbstring_core" &:: "category" &--> Misc.return_int Int64.zero
       ; -"folly" &:: "DelayedDestruction" &:: "destroy" &--> Misc.skip
       ; -"folly" &:: "Optional" &:: "reset" &--> Misc.skip
       ; -"folly" &:: "SocketAddress" &:: "~SocketAddress" &--> Misc.skip