csgrep: chain filters based on the predicate keys

... sorted alphabetically

It would make more sense if the predicates that are cheap to evaluate
(or the predicates that drop more data) were chained first.  On the
other hand, when processing big data sets, the biggest bottleneck is
usually the JSON reader/writer anyway.

PR: #51

Author: kdudka

src/csgrep.cc

@@ -492,13 +492,13 @@ bool chainFiltersCore(
         const po::variables_map                         &vm,
         const TFlags                                    flags)
 {
-    return appendPredIfNeeded<CheckerPredicate>   (pf, vm, flags, "checker")
-        && appendPredIfNeeded<SrcAnnotPredicate>  (pf, vm, flags, "src-annot")
-        && appendPredIfNeeded<AnnotPredicate>     (pf, vm, flags, "annot")
+    return appendPredIfNeeded<AnnotPredicate>     (pf, vm, flags, "annot")
+        && appendPredIfNeeded<CheckerPredicate>   (pf, vm, flags, "checker")
         && appendPredIfNeeded<ErrorPredicate>     (pf, vm, flags, "error")
         && appendPredIfNeeded<KeyEventPredicate>  (pf, vm, flags, "event")
         && appendPredIfNeeded<MsgPredicate>       (pf, vm, flags, "msg")
-        && appendPredIfNeeded<PathPredicate>      (pf, vm, flags, "path");
+        && appendPredIfNeeded<PathPredicate>      (pf, vm, flags, "path")
+        && appendPredIfNeeded<SrcAnnotPredicate>  (pf, vm, flags, "src-annot");
 }
 
 bool chainFilters(