[filtering/dedup] add deduplicate flag and disambiguate filtering

Summary:
The documentation and uses of filtering disagree.  One typical usage is deduplication.
Split that where obvious, add comments where not obvious, and leave alone when obviously unrelated to deduplication.

Reviewed By: mityal

Differential Revision: D17715329

fbshipit-source-id: ec757927b

infer/man/man1/infer-analyze.txt

@@ -96,6 +96,10 @@ OPTIONS
            Debug level for the linters. See --debug-level for accepted
            values.
 
+       --no-deduplicate
+           Deactivates: Apply issue-specific deduplication during analysis
+           and/or reporting. (Conversely: --deduplicate)
+
        --no-default-checkers
            Deactivates: Default checkers: --biabduction,
            --fragment-retains-view, --inefficient-keyset-iterator, --linters,

infer/man/man1/infer-full.txt

@@ -296,6 +296,11 @@ OPTIONS
            infer-events(1), infer-report(1), infer-reportdiff(1), and
            infer-run(1).
 
+       --no-deduplicate
+           Deactivates: Apply issue-specific deduplication during analysis
+           and/or reporting. (Conversely: --deduplicate)           See also infer-analyze(1), infer-report(1), and
+           infer-reportdiff(1).
+
        --no-default-checkers
            Deactivates: Default checkers: --biabduction,
            --fragment-retains-view, --inefficient-keyset-iterator, --linters,
@@ -1277,8 +1282,9 @@ INTERNAL OPTIONS
        --debug-exceptions
            Activates: Generate lightweight debugging information: just print
            the internal exceptions during analysis (also sets
-           --developer-mode, --no-filtering, --print-buckets,
-           --reports-include-ml-loc) (Conversely: --no-debug-exceptions)
+           --developer-mode, --no-filtering, --no-deduplicate,
+           --print-buckets, --reports-include-ml-loc) (Conversely:
+           --no-debug-exceptions)
 
        --debug-level-reset
            Cancel the effect of --debug-level.

infer/man/man1/infer-report.txt

@@ -60,6 +60,10 @@ OPTIONS
            Debug level for the linters. See --debug-level for accepted
            values.
 
+       --no-deduplicate
+           Deactivates: Apply issue-specific deduplication during analysis
+           and/or reporting. (Conversely: --deduplicate)
+
        --differential-filter-files string
            Specify the file containing the list of source files for which a
            differential report is desired. Source files should be specified

infer/man/man1/infer-reportdiff.txt

@@ -51,6 +51,10 @@ OPTIONS
            Debug level for the linters. See --debug-level for accepted
            values.
 
+       --no-deduplicate
+           Deactivates: Apply issue-specific deduplication during analysis
+           and/or reporting. (Conversely: --deduplicate)
+
        --file-renamings path
            JSON with a list of file renamings to use while computing
            differential reports

infer/man/man1/infer.txt

@@ -296,6 +296,11 @@ OPTIONS
            infer-events(1), infer-report(1), infer-reportdiff(1), and
            infer-run(1).
 
+       --no-deduplicate
+           Deactivates: Apply issue-specific deduplication during analysis
+           and/or reporting. (Conversely: --deduplicate)           See also infer-analyze(1), infer-report(1), and
+           infer-reportdiff(1).
+
        --no-default-checkers
            Deactivates: Default checkers: --biabduction,
            --fragment-retains-view, --inefficient-keyset-iterator, --linters,

infer/src/backend/Differential.ml

@@ -19,6 +19,7 @@ module LocListSet = struct
 end
 
 let is_duplicate_report end_locs reported_ends =
+  (* FIXME(T54950303) replace use of filtering with deduplicate *)
   Config.filtering && LocListSet.mem reported_ends end_locs
 
 

infer/src/backend/Issue.ml

@@ -38,6 +38,7 @@ type t_ignore_duplicates = t =
      de-duplicating. *)
 let sort_filter_issues issues =
   let issues' =
+    (* FIXME(T54950303) replace use of filtering with deduplicate *)
     let compare = if Config.filtering then compare_t_ignore_duplicates else compare in
     List.dedup_and_sort ~compare issues
   in

infer/src/base/Config.ml

@@ -1136,6 +1136,7 @@ and custom_symbols =
 
 and ( biabduction_models_mode
     , bo_debug
+    , deduplicate
     , developer_mode
     , debug
     , debug_exceptions
@@ -1168,6 +1169,12 @@ and ( biabduction_models_mode
     CLOpt.mk_int ~default:0 ~long:"bo-debug"
       ~in_help:InferCommand.[(Analyze, manual_buffer_overrun)]
       "Debug level for buffer-overrun checker (0-4)"
+  and deduplicate =
+    CLOpt.mk_bool ~long:"deduplicate" ~default:true
+      ~in_help:
+        InferCommand.
+          [(Analyze, manual_generic); (Report, manual_generic); (ReportDiff, manual_generic)]
+      "Apply issue-specific deduplication during analysis and/or reporting."
   and debug_level_analysis =
     CLOpt.mk_int ~long:"debug-level-analysis" ~default:0 ~in_help:all_generic_manuals
       "Debug level for the analysis. See $(b,--debug-level) for accepted values."
@@ -1248,10 +1255,10 @@ and ( biabduction_models_mode
   and debug_exceptions =
     CLOpt.mk_bool_group ~long:"debug-exceptions"
       "Generate lightweight debugging information: just print the internal exceptions during \
-       analysis (also sets $(b,--developer-mode), $(b,--no-filtering), $(b,--print-buckets), \
-       $(b,--reports-include-ml-loc))"
+       analysis (also sets $(b,--developer-mode), $(b,--no-filtering), $(b,--no-deduplicate), \
+       $(b,--print-buckets), $(b,--reports-include-ml-loc))"
       [developer_mode; print_buckets; reports_include_ml_loc]
-      [filtering; keep_going]
+      [filtering; keep_going; deduplicate]
   and default_linters =
     CLOpt.mk_bool ~long:"default-linters"
       ~in_help:InferCommand.[(Capture, manual_clang_linters)]
@@ -1285,6 +1292,7 @@ and ( biabduction_models_mode
   in
   ( biabduction_models_mode
   , bo_debug
+  , deduplicate
   , developer_mode
   , debug
   , debug_exceptions
@@ -2837,6 +2845,8 @@ and debug_exceptions = !debug_exceptions
 
 and debug_mode = !debug
 
+and deduplicate = !deduplicate
+
 and default_linters = !default_linters
 
 and dependency_mode = !dependencies

infer/src/base/Config.mli

@@ -312,6 +312,8 @@ val cxx : bool
 
 val cxx_scope_guards : Yojson.Basic.t
 
+val deduplicate : bool
+
 val debug_exceptions : bool
 
 val debug_level_analysis : int

infer/src/checkers/Siof.ml

@@ -242,6 +242,7 @@ let report_siof summary trace gname loc =
     Reporting.log_error summary ~loc ~ltr IssueType.static_initialization_order_fiasco description
   in
   let reportable_paths = SiofTrace.get_reportable_sink_paths trace ~trace_of_pname in
+  (* FIXME(T54950303) replace use of filtering with deduplicate *)
   if Config.filtering then List.hd reportable_paths |> Option.iter ~f:report_one_path
   else List.iter ~f:report_one_path reportable_paths
 

infer/src/concurrency/RacerD.ml

@@ -1000,7 +1000,7 @@ let report_unsafe_accesses ~issue_log classname (aggregated_access_map : ReportM
       ({reported_sites; reported_writes; reported_reads; reported_unannotated_calls}, _) =
     let pname = Procdesc.get_proc_name procdesc in
     let call_site = CallSite.make pname (TraceElem.get_loc snapshot.access) in
-    if Config.filtering then
+    if Config.deduplicate then
       CallSite.Set.mem call_site reported_sites
       ||
       match snapshot.access.TraceElem.elem with
@@ -1013,7 +1013,7 @@ let report_unsafe_accesses ~issue_log classname (aggregated_access_map : ReportM
     else false
   in
   let update_reported ({snapshot; procdesc} : reported_access) reported =
-    if Config.filtering then
+    if Config.deduplicate then
       let pname = Procdesc.get_proc_name procdesc in
       let call_site = CallSite.make pname (TraceElem.get_loc snapshot.access) in
       let reported_sites = CallSite.Set.add call_site reported.reported_sites in

infer/src/integration/ReportDiff.ml

@@ -21,6 +21,7 @@ let reportdiff ~current_report:current_report_fname ~previous_report:previous_re
     let unfiltered_diff =
       Differential.of_reports ~current_report ~previous_report ~current_costs ~previous_costs
     in
+    (* FIXME(T54950303) replace use of filtering with deduplicate *)
     if Config.filtering then
       let file_renamings =
         match Config.file_renamings with

infer/src/nullsafe/NullabilityCheck.ml

@@ -73,6 +73,7 @@ module TransferFunctions (CFG : ProcCfg.S) = struct
 
 
   let is_conflicting_report summary report_location =
+    (* FIXME(T54950303) replace use of filtering with deduplicate *)
     if not Config.filtering then false
     else
       Errlog.fold