diagnostics: suggest fixes for undeclared symbols

From

    input.y:1.17-19: warning: symbol baz is used, but is not defined as a token and has no rules [-Wother]
         1 | %printer {} foo baz
           |                 ^~~

to

    input.y:1.17-19: warning: symbol 'baz' is used, but is not defined as a token and has no rules; did you mean 'bar'? [-Wother]
        1 | %printer {} foo baz
          |                 ^~~
          |                 bar

* bootstrap.conf: We need fstrcmp.
* src/symtab.c (symbol_from_uniqstr_fuzzy): New.
(complain_symbol_undeclared): Use it.
* tests/diagnostics.at (Suggestions): New.
* data/bison-default.css (insertion): Rename as...
(fixit-insert): this, as this is what GCC uses.

bootstrap.conf

@@ -22,7 +22,8 @@ gnulib_modules='
   calloc-posix close closeout config-h c-strcase
   configmake
   dirname
-  error extensions fdl fopen-safer
+  error extensions
+  fdl fopen-safer fstrcmp
   getopt-gnu
   gettext-h git-version-gen gitlog-to-changelog
   gpl-3.0 intprops inttypes isnan javacomp-script

data/bison-default.css

@@ -21,7 +21,8 @@
 .warning   { color: purple; }
 .error     { color: red; }
 .note      { color: cyan; }
-.insertion { color: green; }
+
+.fixit-insert { color: green; }
 
 /* Semantic values in Bison's own parser traces.  */
 .value     { color: green; }

lib/.gitignore

@@ -53,6 +53,7 @@
 /config.h
 /config.in.h
 /configmake.h
+/diffseq.h
 /dirname-lgpl.c
 /dirname.c
 /dirname.h
@@ -91,6 +92,8 @@
 /fseterr.c
 /fseterr.h
 /fstat.c
+/fstrcmp.c
+/fstrcmp.h
 /fsync.c
 /getdtablesize.c
 /gethrxtime.c
@@ -324,6 +327,8 @@
 /windows-recmutex.h
 /windows-rwlock.c
 /windows-rwlock.h
+/windows-tls.c
+/windows-tls.h
 /xalloc-die.c
 /xalloc-oversized.h
 /xalloc.h

lib/glthread/.gitignore

@@ -1,3 +1,5 @@
 /lock.c
 /lock.h
 /threadlib.c
+/tls.c
+/tls.h

m4/.gitignore

@@ -71,6 +71,8 @@
 /iswblank.m4
 /javacomp.m4
 /javaexec.m4
+/jm-winsz1.m4
+/jm-winsz2.m4
 /largefile.m4
 /lcmessage.m4
 /ldexp.m4
@@ -179,6 +181,7 @@
 /threadlib.m4
 /time_h.m4
 /timespec.m4
+/tls.m4
 /uintmax_t.m4
 /unistd-safer.m4
 /unistd_h.m4
@@ -202,5 +205,3 @@
 /xalloc.m4
 /xsize.m4
 /xstrndup.m4
-/jm-winsz1.m4
-/jm-winsz2.m4

src/location.c

@@ -473,7 +473,7 @@ location_caret (location loc, const char *style, FILE *out)
 void
 location_caret_suggestion (location loc, const char *s, FILE *out)
 {
-  const char *style = "insertion";
+  const char *style = "fixit-insert";
   fprintf (out, "      | %*s",
            loc.start.column - 1 - caret_info.skip + (caret_info.skip ? 3 : 0),
            "");

src/symtab.c

@@ -24,6 +24,7 @@
 #include "system.h"
 
 #include <assure.h>
+#include <fstrcmp.h>
 #include <hash.h>
 
 #include "complain.h"
@@ -32,6 +33,8 @@
 #include "quote.h"
 
 
+static struct hash_table *symbol_table = NULL;
+static struct hash_table *semantic_type_table = NULL;
 
 /*----------------------------------------------------------------.
 | Symbols sorted by tag.  Allocated by table_sort, after which no |
@@ -41,6 +44,7 @@
 static symbol **symbols_sorted = NULL;
 static semantic_type **semantic_types_sorted = NULL;
 
+
 /*------------------------.
 | Distinguished symbols.  |
 `------------------------*/
@@ -321,15 +325,54 @@ complain_class_redeclared (symbol *sym, symbol_class class, location second)
     }
 }
 
+static const symbol *
+symbol_from_uniqstr_fuzzy (const uniqstr key)
+{
+  aver (symbols_sorted);
+#define FSTRCMP_THRESHOLD 0.6
+  double best_similarity = FSTRCMP_THRESHOLD;
+  const symbol *res = NULL;
+  size_t count = hash_get_n_entries (symbol_table);
+  for (int i = 0; i < count; ++i)
+    {
+      symbol *sym = symbols_sorted[i];
+      if (STRNEQ (key, sym->tag)
+          && (sym->content->status == declared
+              || sym->content->status == undeclared))
+        {
+          double similarity = fstrcmp_bounded (key, sym->tag, best_similarity);
+          if (best_similarity < similarity)
+            {
+              res = sym;
+              best_similarity = similarity;
+            }
+        }
+    }
+  return res;
+}
+
 static void
 complain_symbol_undeclared (symbol *sym)
 {
   assert (sym->content->status != declared);
-  complain (&sym->location,
+  const symbol *best = symbol_from_uniqstr_fuzzy (sym->tag);
-            sym->content->status == needed ? complaint : Wother,
+  if (best)
-            _("symbol %s is used, but is not defined as a token"
+    {
-              " and has no rules"),
+      complain (&sym->location,
-            quote (sym->tag));
+                sym->content->status == needed ? complaint : Wother,
+                _("symbol %s is used, but is not defined as a token"
+                  " and has no rules; did you mean %s?"),
+                quote_n (0, sym->tag),
+                quote_n (1, best->tag));
+      if (feature_flag & feature_caret)
+        location_caret_suggestion (sym->location, best->tag, stderr);
+    }
+  else
+    complain (&sym->location,
+              sym->content->status == needed ? complaint : Wother,
+              _("symbol %s is used, but is not defined as a token"
+                " and has no rules"),
+              quote (sym->tag));
 }
 
 void
@@ -694,9 +737,6 @@ symbol_translation (symbol *this)
 /* Initial capacity of symbol and semantic type hash table.  */
 #define HT_INITIAL_CAPACITY 257
 
-static struct hash_table *symbol_table = NULL;
-static struct hash_table *semantic_type_table = NULL;
-
 static inline bool
 hash_compare_symbol (const symbol *m1, const symbol *m2)
 {

tests/diagnostics.at

@@ -46,7 +46,7 @@ AT_DATA([experr], [$4])
 AT_CHECK([LC_ALL=en_US.UTF-8 $5 bison -fcaret --color=debug -Wall input.y], [$3], [], [experr])
 
 # When no style, same messages, but without style.
-AT_CHECK([perl -pi -e 's{(</?\w+>)}{ $[]1 eq "<tag>" ? $[]1 : "" }ge' experr])
+AT_CHECK([perl -pi -e 's{(</?(-|\w)+>)}{ $[]1 eq "<tag>" ? $[]1 : "" }ge' experr])
 
 # Cannot use AT_BISON_CHECK easily as we need to change the
 # environment.
@@ -344,7 +344,7 @@ input.y:9.8-33:       previous declaration
     9 | %token <note>ABCDEFGHIJKLMNOPQRSTUVWXYZ</note>  ABCDEFGHIJKLMNOPQRSTUVWXYZ  ABCDEFGHIJKLMNOPQRSTUVWXYZ  ABCDEFGHIJKLMNOPQRSTUVWXYZ
       |        <note>^~~~~~~~~~~~~~~~~~~~~~~~~~</note>
 input.y:10.56-69: <warning>warning:</warning> deprecated directive: '%error-verbose', use '%define parse.error verbose' [<warning>-Wdeprecated</warning>]
-   10 |                                                        <warning>%error-verbose</warning>
+   10 |                                                        <warning>%error-verbose</warning>
       |                                                        <warning>^~~~~~~~~~~~~~</warning>
       |                                                        <fixit-insert>%define parse.error verbose</fixit-insert>
 input.y: <warning>warning:</warning> fix-its can be applied.  Rerun with option '--update'. [<warning>-Wother</warning>]
@@ -377,7 +377,7 @@ input.y:9.8-33:       previous declaration
     9 | %token <note>ABCDEFGHIJKLMNOPQRSTUVWXYZ</note>  ABCDEFGHIJKLMNOPQRSTUVWXYZ  ABCDEF...
       |        <note>^~~~~~~~~~~~~~~~~~~~~~~~~~</note>
 input.y:10.56-69: <warning>warning:</warning> deprecated directive: '%error-verbose', use '%define parse.error verbose' [<warning>-Wdeprecated</warning>]
-   10 |                                                        <warning>%error-verbose</warning>
+   10 |                                                        <warning>%error-verbose</warning>
       |                                                        <warning>^~~~~~~~~~~~~~</warning>
       |                                                        <fixit-insert>%define parse.error verbose</fixit-insert>
 input.y: <warning>warning:</warning> fix-its can be applied.  Rerun with option '--update'. [<warning>-Wother</warning>]
@@ -410,11 +410,39 @@ input.y:9.8-33:       previous declaration
     9 | %token <note>ABCDEFGHIJKLMNOPQRSTUVWXYZ</note>  ABCDEFGHIJKLMN...
       |        <note>^~~~~~~~~~~~~~~~~~~~~~~~~~</note>
 input.y:10.56-69: <warning>warning:</warning> deprecated directive: '%error-verbose', use '%define parse.error verbose' [<warning>-Wdeprecated</warning>]
-   10 | ...         <warning>%error-verbose</warning>
+   10 | ...         <warning>%error-verbose</warning>
       |             <warning>^~~~~~~~~~~~~~</warning>
       |             <fixit-insert>%define parse.error verbose</fixit-insert>
 input.y: <warning>warning:</warning> fix-its can be applied.  Rerun with option '--update'. [<warning>-Wother</warning>]
 ]],
+[[COLUMNS=60]])
+
+
+## ------------- ##
+## Suggestions.  ##
+## ------------- ##
+
+# Don't suggest to fix QUX with QUUX and QUUX with QUX...
+AT_TEST([[Suggestions]],
+[[%%
+res: QUX baz
+bar: QUUX
+]],
+[1],
+[[input.y:11.6-9: <error>error:</error> symbol 'QUUX' is used, but is not defined as a token and has no rules
+   11 | bar: <error>QUUX</error>
+      |      <error>^~~~</error>
+input.y:10.6-8: <error>error:</error> symbol 'QUX' is used, but is not defined as a token and has no rules
+   10 | res: <error>QUX</error> baz
+      |      <error>^~~</error>
+input.y:10.10-12: <error>error:</error> symbol 'baz' is used, but is not defined as a token and has no rules; did you mean 'bar'?
+   10 | res: QUX <error>baz</error>
+      |          <error>^~~</error>
+      |          <fixit-insert>bar</fixit-insert>
+]])
+
+
+
 
 
 m4_popdef([AT_TEST])