[Bug other/94982] New: '-Wformat-diag' diagnostics building GCC

[Bug other/94982] New: '-Wformat-diag' diagnostics building GCC

[tschwinge at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=94982

            Bug ID: 94982
           Summary: '-Wformat-diag' diagnostics building GCC
           Product: gcc
           Version: 10.0
            Status: UNCONFIRMED
          Keywords: diagnostic
          Severity: normal
          Priority: P3
         Component: other
          Assignee: unassigned at gcc dot gnu.org
          Reporter: tschwinge at gcc dot gnu.org
                CC: msebor at gcc dot gnu.org
  Target Milestone: ---

In a x86_64-pc-linux-gnu GCC bootstrap build of commit
c416c52bcdb120db5e8c53a51bd78c4360daf79b (releases/gcc-10 branch point), I'm
counting 642 'warning: .* \[-Wformat-diag\]' diagnostics.  For some files it's
hard to spot other diagnostics between all that "noise".  I'm not saying that
'-Wformat-diag' diagnostics don't have any value (they do!), but there are
clearly bogus ones here -- we should do something about this, or disable if not
interested?

[Bug other/94982] '-Wformat-diag' diagnostics building GCC

[msebor at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=94982

Martin Sebor <msebor at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |dmalcolm at gcc dot gnu.org
   Last reconfirmed|                            |2020-05-07
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |NEW

--- Comment #1 from Martin Sebor <msebor at gcc dot gnu.org> ---
The breakdown of the -Wformat-diag warnings from a recent x86_64-linux build of
mine is below.  That's about 200.  Where do all the others come from in your
build?

The largest subset of them are coming from the analyzer.  Those I've looked at
are either for calls to logging functions, or to pp_xxx functions that don't
follow the expected format of GCC diagnostics.  I'm not familiar enough with
the analyzer to say what the right way to deal with all those is.  The messages
that are meant to be conform to the same formatting and spelling rules as
diagnostics should be cleaned up.  Those that are more like the output of
-ftree-dump-xxx options should be suppressed (e.g., by a #pragma or by adding a
new attribute to exempt them from the warning).

The second largest subset is in gcc/cp/error.c.  Those are caused by calls to
the pp_xxx functions that compose messages one piece at a time, with the
individual pieces not conforming to the expected format of complete diagnostic
messages.  These should be suppressed (again, either by a #pragma or by adding
a new attribute).

I haven't looked craefully at the rest but I suspect they're in the pp_xxx
category.

My hope and expectation when I introduced the warnings (and kept them from
causing errors) was that they would get cleaned up over the course of GCC 10
development by contributors to and maintainers of the affected areas. 
Unfortunately, not only has that not happened but the error suppression made it
possible for regressions to creep in.

Let me work on cleaning up those outside the analyzer and outside of back ends.
 I'll leave the analyzer to David.  I'm hoping back end maintainers will have
dealt with any warnings there by now however they feel is appropriate.  Once
the remaining instances are all cleaned up I'd like to remove the -Werror
suppression for -Wformat-diag to keep any more regressions from sneaking in
(that's always been the goal).

-Wformat-diag Instances:
  gcc/analyzer/analyzer-logging.cc:174
  gcc/analyzer/analyzer-logging.cc:79
  gcc/analyzer/call-string.cc:97
  gcc/analyzer/checker-path.cc:112
  gcc/analyzer/checker-path.cc:902
  gcc/analyzer/checker-path.cc:916
  gcc/analyzer/checker-path.cc:922
  gcc/analyzer/constraint-manager.cc:574
  gcc/analyzer/constraint-manager.cc:601
  gcc/analyzer/diagnostic-manager.cc:381
  gcc/analyzer/diagnostic-manager.cc:391
  gcc/analyzer/diagnostic-manager.cc:402
  gcc/analyzer/diagnostic-manager.cc:414
  gcc/analyzer/diagnostic-manager.cc:432
  gcc/analyzer/diagnostic-manager.cc:468
  gcc/analyzer/diagnostic-manager.cc:806
  gcc/analyzer/diagnostic-manager.cc:940
  gcc/analyzer/diagnostic-manager.cc:942
  gcc/analyzer/engine.cc:1496
  gcc/analyzer/engine.cc:1508
  gcc/analyzer/engine.cc:1532
  gcc/analyzer/engine.cc:1535
  gcc/analyzer/engine.cc:1536
  gcc/analyzer/engine.cc:2229
  gcc/analyzer/engine.cc:253
  gcc/analyzer/engine.cc:2557
  gcc/analyzer/engine.cc:2573
  gcc/analyzer/engine.cc:2728
  gcc/analyzer/engine.cc:2729
  gcc/analyzer/engine.cc:2730
  gcc/analyzer/engine.cc:2847
  gcc/analyzer/engine.cc:2877
  gcc/analyzer/engine.cc:2932
  gcc/analyzer/engine.cc:2953
  gcc/analyzer/engine.cc:3013
  gcc/analyzer/engine.cc:3016
  gcc/analyzer/engine.cc:3017
  gcc/analyzer/engine.cc:3027
  gcc/analyzer/engine.cc:3061
  gcc/analyzer/engine.cc:3064
  gcc/analyzer/engine.cc:3066
  gcc/analyzer/engine.cc:3067
  gcc/analyzer/engine.cc:3076
  gcc/analyzer/engine.cc:3446
  gcc/analyzer/engine.cc:3509
  gcc/analyzer/engine.cc:3520
  gcc/analyzer/engine.cc:3525
  gcc/analyzer/engine.cc:3540
  gcc/analyzer/engine.cc:3565
  gcc/analyzer/engine.cc:3577
  gcc/analyzer/engine.cc:3589
  gcc/analyzer/engine.cc:3622
  gcc/analyzer/engine.cc:3625
  gcc/analyzer/engine.cc:3769
  gcc/analyzer/engine.cc:389
  gcc/analyzer/engine.cc:416
  gcc/analyzer/engine.cc:572
  gcc/analyzer/engine.cc:846
  gcc/analyzer/engine.cc:871
  gcc/analyzer/engine.cc:903
  gcc/analyzer/program-point.cc:120
  gcc/analyzer/program-state.cc:1072
  gcc/analyzer/program-state.cc:173
  gcc/analyzer/program-state.cc:688
  gcc/analyzer/program-state.cc:69
  gcc/analyzer/program-state.cc:698
  gcc/analyzer/program-state.cc:717
  gcc/analyzer/program-state.cc:726
  gcc/analyzer/program-state.cc:74
  gcc/analyzer/program-state.cc:806
  gcc/analyzer/program-state.cc:928
  gcc/analyzer/region-model.cc:1246
  gcc/analyzer/region-model.cc:1300
  gcc/analyzer/region-model.cc:1306
  gcc/analyzer/region-model.cc:1488
  gcc/analyzer/region-model.cc:1493
  gcc/analyzer/region-model.cc:1598
  gcc/analyzer/region-model.cc:1636
  gcc/analyzer/region-model.cc:1643
  gcc/analyzer/region-model.cc:1669
  gcc/analyzer/region-model.cc:1671
  gcc/analyzer/region-model.cc:193
  gcc/analyzer/region-model.cc:2132
  gcc/analyzer/region-model.cc:223
  gcc/analyzer/region-model.cc:2282
  gcc/analyzer/region-model.cc:2320
  gcc/analyzer/region-model.cc:2327
  gcc/analyzer/region-model.cc:2352
  gcc/analyzer/region-model.cc:2576
  gcc/analyzer/region-model.cc:2902
  gcc/analyzer/region-model.cc:2904
  gcc/analyzer/region-model.cc:2906
  gcc/analyzer/region-model.cc:2908
  gcc/analyzer/region-model.cc:3583
  gcc/analyzer/region-model.cc:360
  gcc/analyzer/region-model.cc:3651
  gcc/analyzer/region-model.cc:3660
  gcc/analyzer/region-model.cc:3741
  gcc/analyzer/region-model.cc:6364
  gcc/analyzer/sm.cc:103
  gcc/analyzer/sm-file.cc:379
  gcc/analyzer/sm-file.cc:385
  gcc/analyzer/sm-malloc.cc:789
  gcc/analyzer/sm-malloc.cc:795
  gcc/analyzer/state-purge.cc:117
  gcc/analyzer/state-purge.cc:135
  gcc/analyzer/state-purge.cc:185
  gcc/analyzer/state-purge.cc:193
  gcc/analyzer/state-purge.cc:216
  gcc/analyzer/state-purge.cc:258
  gcc/analyzer/state-purge.cc:260
  gcc/analyzer/state-purge.cc:295
  gcc/analyzer/state-purge.cc:297
  gcc/analyzer/state-purge.cc:323
  gcc/analyzer/state-purge.cc:361
  gcc/analyzer/state-purge.cc:431
  gcc/analyzer/state-purge.cc:432
  gcc/analyzer/state-purge.cc:473
  gcc/analyzer/state-purge.cc:480
  gcc/analyzer/supergraph.cc:154
  gcc/analyzer/supergraph.cc:270
  gcc/analyzer/supergraph.cc:271
  gcc/analyzer/supergraph.cc:284
  gcc/analyzer/supergraph.cc:288
  gcc/analyzer/supergraph.cc:300
  gcc/analyzer/supergraph.cc:304
  gcc/analyzer/supergraph.cc:321
  gcc/analyzer/supergraph.cc:335
  gcc/analyzer/supergraph.cc:347
  gcc/analyzer/supergraph.cc:449
  gcc/analyzer/supergraph.cc:453
  gcc/analyzer/supergraph.cc:454
  gcc/analyzer/supergraph.cc:455
  gcc/analyzer/supergraph.cc:456
  gcc/analyzer/supergraph.cc:466
  gcc/analyzer/supergraph.cc:552
  gcc/analyzer/supergraph.cc:560
  gcc/analyzer/supergraph.cc:632
  gcc/analyzer/supergraph.cc:711
  gcc/analyzer/supergraph.cc:719
  gcc/analyzer/supergraph.cc:828
  gcc/analyzer/supergraph.cc:832
  gcc/analyzer/supergraph.cc:835
  gcc/cgraphunit.c:763
  gcc/cp/constraint.cc:535
  gcc/cp/constraint.cc:537
  gcc/cp/error.c:3496
  gcc/cp/error.c:3497
  gcc/cp/error.c:3522
  gcc/cp/error.c:3525
  gcc/cp/error.c:3533
  gcc/cp/error.c:3534
  gcc/cp/error.c:3539
  gcc/cp/error.c:3540
  gcc/cp/error.c:3547
  gcc/cp/error.c:3548
  gcc/cp/error.c:3600
  gcc/cp/error.c:3606
  gcc/cp/error.c:3660
  gcc/cp/error.c:3664
  gcc/cp/error.c:3676
  gcc/cp/error.c:3679
  gcc/cp/error.c:3708
  gcc/cp/error.c:3725
  gcc/cp/error.c:3738
  gcc/cp/error.c:3765
  gcc/cp/error.c:3769
  gcc/cp/error.c:3786
  gcc/cp/error.c:3946
  gcc/cp/error.c:4004
  gcc/cp/error.c:4108
  gcc/cp/error.c:768
  gcc/cp/logic.cc:321
  gcc/cp/logic.cc:323
  gcc/cp/parser.c:29138
  gcc/cp/parser.c:9239
  gcc/cp/typeck.c:2879
  gcc/dbgcnt.c:129
  gcc/digraph.cc:70
  gcc/dumpfile.c:1120
  gcc/dumpfile.c:496
  gcc/dumpfile.c:499
  gcc/dumpfile.c:503
  gcc/edit-context.c:450
  gcc/edit-context.c:451
  gcc/edit-context.c:522
  gcc/jit/jit-recording.c:3941
  gcc/jit/jit-recording.c:3959
  gcc/jit/jit-recording.c:4347
  gcc/jit/jit-recording.c:4369
  gcc/jit/jit-recording.c:4383
  gcc/json.cc:73
  gcc/lto-wrapper.c:504
  gcc/lto-wrapper.c:505
  gcc/lto-wrapper.c:512
  gcc/lto-wrapper.c:513
  gcc/lto-wrapper.c:521
  gcc/pretty-print.c:2179
  gcc/pretty-print.c:2182
  gcc/tree-diagnostic-path.cc:419
  gcc/tree-diagnostic-path.cc:627
  gcc/ubsan.c:409
  gcc/ubsan.c:428

[Bug other/94982] '-Wformat-diag' diagnostics building GCC

[msebor at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=94982

Martin Sebor <msebor at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |patch

--- Comment #2 from Martin Sebor <msebor at gcc dot gnu.org> ---
Patch: https://gcc.gnu.org/pipermail/gcc-patches/2020-November/560129.html

[Bug other/94982] '-Wformat-diag' diagnostics building GCC

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=94982

--- Comment #3 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Martin Sebor <msebor@gcc.gnu.org>:

https://gcc.gnu.org/g:ca23341b28cd3af7985b83a6479107d9ea145a90

commit r11-5393-gca23341b28cd3af7985b83a6479107d9ea145a90
Author: Martin Sebor <msebor@redhat.com>
Date:   Wed Nov 25 14:05:01 2020 -0700

    Clean up -Wformat-diag warnings (PR bootstrap/97622, PR bootstrap/94982)

    gcc/c-family/ChangeLog:

            PR bootstrap/94982
            * c-attribs.c (handle_patchable_function_entry_attribute): Avoid
            -Wformat-diag.

    gcc/cp/ChangeLog:

            PR bootstrap/94982
            * constraint.cc (debug_argument_list): Avoid -Wformat-diag.
            * error.c (function_category): Same.
            (print_template_differences): Same.
            * logic.cc (debug): Same.
            * name-lookup.c (lookup_using_decl): Same.
            * parser.c (maybe_add_cast_fixit): Same.
            (cp_parser_template_introduction): Same.
            * typeck.c (access_failure_info::add_fixit_hint): Same.

    gcc/ChangeLog:

            PR bootstrap/97622
            PR bootstrap/94982
            * config/i386/i386-options.c (ix86_valid_target_attribute_inner_p):
            Avoid -Wformat-diag.
            * digraph.cc (struct test_edge): Same.
            * dumpfile.c (dump_loc): Same.
            (dump_context::begin_scope): Same.
            * edit-context.c (edited_file::print_diff): Same.
            (edited_file::print_diff_hunk): Same.
            * json.cc (object::print): Same.
            * lto-wrapper.c (merge_and_complain): Same.
            * reload.c (find_reloads): Same.
            * tree-diagnostic-path.cc (print_path_summary_as_text): Same.
            * ubsan.c (ubsan_type_descriptor): Same.

    gcc/jit/ChangeLog:

            PR bootstrap/94982
            * jit-recording.c (recording::function::dump_to_dot): Avoid
            -Wformat-diag.
            (recording::block::dump_to_dot): Same.

    gcc/testsuite/ChangeLog:

            PR bootstrap/94982
            * c-c++-common/patchable_function_entry-error-3.c: Adjust text
            of expected warning.

[Bug other/94982] '-Wformat-diag' diagnostics building GCC

[msebor at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=94982

Martin Sebor <msebor at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |FIXED
             Status|NEW                         |RESOLVED

--- Comment #4 from Martin Sebor <msebor at gcc dot gnu.org> ---
With the patch committed in r11-5393 an x86_64-linux build should be free of
most if not all -Wformat-diag instances.  I'm going to resolve this as fixed. 
Let's track any outstanding instances or warnings on other targets as separate
bugs.