Change 29502 wasn't perfect - you need to remove any extra trailing

"\n" added by fbm_compile(), before recompiling with the same flags.
In turn, to do that, it's best to store the flags even for short
"PVBM"s.

p4raw-id: //depot/perl@30092

diff --git a/regexec.c b/regexec.c
index 6c82ba7..8697eb6 100644 (file)
--- a/regexec.c
+++ b/regexec.c
@@ -5690,10 +5690,18 @@ S_to_utf8_substr(pTHX_ register regexp *prog)
            SV* const sv = newSVsv(prog->substrs->data[i].substr);
            prog->substrs->data[i].utf8_substr = sv;
            sv_utf8_upgrade(sv);
-           if (SvVALID(prog->substrs->data[i].substr))
-               fbm_compile(sv, 0);
-           if (SvTAIL(prog->substrs->data[i].substr))
-               SvTAIL_on(sv);
+           if (SvVALID(prog->substrs->data[i].substr)) {
+               const U8 flags = BmFLAGS(prog->substrs->data[i].substr);
+               if (flags & FBMcf_TAIL) {
+                   /* Trim the trailing \n that fbm_compile added last
+                      time.  */
+                   SvCUR_set(sv, SvCUR(sv) - 1);
+                   /* Whilst this makes the SV technically "invalid" (as its
+                      buffer is no longer followed by "\0") when fbm_compile()
+                      adds the "\n" back, a "\0" is restored.  */
+               }
+               fbm_compile(sv, flags);
+           }
            if (prog->substrs->data[i].substr == prog->check_substr)
                prog->check_utf8 = sv;
        }
@@ -5710,10 +5718,16 @@ S_to_byte_substr(pTHX_ register regexp *prog)
            && !prog->substrs->data[i].substr) {
            SV* sv = newSVsv(prog->substrs->data[i].utf8_substr);
            if (sv_utf8_downgrade(sv, TRUE)) {
-               if (SvVALID(prog->substrs->data[i].utf8_substr))
-                   fbm_compile(sv, 0);
-               if (SvTAIL(prog->substrs->data[i].utf8_substr))
-                   SvTAIL_on(sv);
+               if (SvVALID(prog->substrs->data[i].utf8_substr)) {
+                   const U8 flags
+                       = BmFLAGS(prog->substrs->data[i].utf8_substr);
+                   if (flags & FBMcf_TAIL) {
+                       /* Trim the trailing \n that fbm_compile added last
+                          time.  */
+                       SvCUR_set(sv, SvCUR(sv) - 1);
+                   }
+                   fbm_compile(sv, flags);
+               }           
            } else {
                SvREFCNT_dec(sv);
                sv = &PL_sv_undef;

diff --git a/util.c b/util.c
index 1e85eca..a8b83f7 100644 (file)
--- a/util.c
+++ b/util.c
@@ -504,7 +504,6 @@ Perl_fbm_compile(pTHX_ SV *sv, U32 flags)
            = (unsigned char*)(SvPVX_mutable(sv) + len + PERL_FBM_TABLE_OFFSET);
        s = table - 1 - PERL_FBM_TABLE_OFFSET;  /* last char */
        memset((void*)table, mlen, 256);
-       BmFLAGS(sv) = (U8)flags;
        i = 0;
        sb = s - mlen + 1;                      /* first char (maybe) */
        while (s >= sb) {
@@ -524,6 +523,7 @@ Perl_fbm_compile(pTHX_ SV *sv, U32 flags)
            frequency = PL_freq[s[i]];
        }
     }
+    BmFLAGS(sv) = (U8)flags;
     BmRARE(sv) = s[rarest];
     BmPREVIOUS(sv) = rarest;
     BmUSEFUL(sv) = 100;                        /* Initial value */