From e4faa4ba4acdd0f04d496b4a6ce5c3260b419f6f Mon Sep 17 00:00:00 2001
From: Karl Williamson <khw@cpan.org>
Date: Sat, 10 Apr 2021 13:32:00 -0600
Subject: [PATCH] locale.c/inline.h: Add fold calc debug statements

Under verbose debugging, this shows non-standard folds
---
 inline.h |  8 +++++-
 locale.c | 84 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 91 insertions(+), 1 deletion(-)

diff --git a/inline.h b/inline.h
index 9c494eba62b7..18ae2ef71cbf 100644
--- a/inline.h
+++ b/inline.h
@@ -2573,8 +2573,14 @@ Perl_foldEQ_locale(const char *s1, const char *s2, I32 len)
     assert(len >= 0);
 
     while (len--) {
-        if (*a != *b && *a != PL_fold_locale[*b])
+        if (*a != *b && *a != PL_fold_locale[*b]) {
+            DEBUG_L(PerlIO_printf(Perl_debug_log,
+                    "%s:%d: Our records indicate %02x is not a fold of %02x"
+                    " or its mate %02x\n",
+                    __FILE__, __LINE__, *a, *b, PL_fold_locale[*b]));
+
             return 0;
+        }
         a++,b++;
     }
     return 1;
diff --git a/locale.c b/locale.c
index 3c2c45559a4e..a339850c3db2 100644
--- a/locale.c
+++ b/locale.c
@@ -1933,6 +1933,27 @@ S_new_ctype(pTHX_ const char *newctype)
         }
     }
     else {  /* Not a canned locale we know the values for.  Compute them */
+
+#    ifdef DEBUGGING
+
+        bool has_non_ascii_fold = FALSE;
+        bool found_unexpected = FALSE;
+
+        if (DEBUG_Lv_TEST) {
+            for (i = 128; i < 256; i++) {
+                if (   toU8_LOWER_LC(LATIN1_TO_NATIVE(i))
+                                  != LATIN1_TO_NATIVE(i)
+                    || toU8_UPPER_LC(LATIN1_TO_NATIVE(i))
+                                  != LATIN1_TO_NATIVE(i))
+                {
+                    has_non_ascii_fold = TRUE;
+                    break;
+                }
+            }
+        }
+
+#    endif
+
         for (i = 0; i < 256; i++) {
             if (isU8_UPPER_LC(i))
                 PL_fold_locale[i] = (U8) toU8_LOWER_LC(i);
@@ -1940,6 +1961,69 @@ S_new_ctype(pTHX_ const char *newctype)
                 PL_fold_locale[i] = (U8) toU8_UPPER_LC(i);
             else
                 PL_fold_locale[i] = (U8) i;
+
+#    ifdef DEBUGGING
+
+            if (DEBUG_Lv_TEST) {
+                bool unexpected = FALSE;
+
+                if (isUPPER_L1(i)) {
+                    if (isUPPER_A(i)) {
+                        if (PL_fold_locale[i] != toLOWER_A(i)) {
+                            unexpected = TRUE;
+                        }
+                    }
+                    else if (has_non_ascii_fold) {
+                        if (PL_fold_locale[i] != toLOWER_L1(i)) {
+                            unexpected = TRUE;
+                        }
+                    }
+                    else if (PL_fold_locale[i] != i) {
+                        unexpected = TRUE;
+                    }
+                }
+                else if (   isLOWER_L1(i)
+                         && i != LATIN_SMALL_LETTER_SHARP_S
+                         && i != MICRO_SIGN)
+                {
+                    if (isLOWER_A(i)) {
+                        if (PL_fold_locale[i] != toUPPER_A(i)) {
+                            unexpected = TRUE;
+                        }
+                    }
+                    else if (has_non_ascii_fold) {
+                        if (PL_fold_locale[i] != toUPPER_LATIN1_MOD(i)) {
+                            unexpected = TRUE;
+                        }
+                    }
+                    else if (PL_fold_locale[i] != i) {
+                        unexpected = TRUE;
+                    }
+                }
+                else if (PL_fold_locale[i] != i) {
+                    unexpected = TRUE;
+                }
+
+                if (unexpected) {
+                    found_unexpected = TRUE;
+                    DEBUG_Lv(PerlIO_printf(Perl_debug_log,
+                                           "For %s, fold of %02x is %02x\n",
+                                           newctype, i, PL_fold_locale[i]));
+                }
+            }
+        }
+
+        if (found_unexpected) {
+            DEBUG_Lv(PerlIO_printf(Perl_debug_log,
+                               "All bytes not mentioned above either fold to"
+                               " themselves or are the expected ASCII or"
+                               " Latin1 ones\n"));
+        }
+        else {
+            DEBUG_Lv(PerlIO_printf(Perl_debug_log,
+                                   "No nonstandard folds were found\n"));
+#    endif
+
         }
     }