5.12.0-RC stuffing bug #10273
Labels
Comments
From zefram@fysh.orgThere's a small bug in lex_stuff_pvn() that causes spurious syntax errors Obviously not a regression from 5.10, since the stuffing API was only -zefram |
From zefram@fysh.orgInline Patchdiff --git a/MANIFEST b/MANIFEST
index a9da994..1f1ce18 100644
--- a/MANIFEST
+++ b/MANIFEST
@@ -3218,6 +3218,7 @@ ext/XS-APItest-KeywordRPN/Makefile.PL XS::APItest::KeywordRPN extension
ext/XS-APItest-KeywordRPN/README XS::APItest::KeywordRPN extension
ext/XS-APItest-KeywordRPN/t/keyword_plugin.t test keyword plugin mechanism
ext/XS-APItest-KeywordRPN/t/multiline.t test plugin parsing across lines
+ext/XS-APItest-KeywordRPN/t/stuff_svcur_bug.t test for a bug in lex_stuff_pvn
ext/XS-APItest/Makefile.PL XS::APItest extension
ext/XS-APItest/MANIFEST XS::APItest extension
ext/XS-APItest/notcore.c Test API functions when PERL_CORE is not defined
diff --git a/ext/XS-APItest-KeywordRPN/KeywordRPN.xs b/ext/XS-APItest-KeywordRPN/KeywordRPN.xs
index e205eea..adb7e6b 100644
--- a/ext/XS-APItest-KeywordRPN/KeywordRPN.xs
+++ b/ext/XS-APItest-KeywordRPN/KeywordRPN.xs
@@ -8,7 +8,7 @@
(!sv_is_glob(sv) && !sv_is_regexp(sv) && \
(SvFLAGS(sv) & (SVf_IOK|SVf_NOK|SVf_POK|SVp_IOK|SVp_NOK|SVp_POK)))
-static SV *hintkey_rpn_sv, *hintkey_calcrpn_sv;
+static SV *hintkey_rpn_sv, *hintkey_calcrpn_sv, *hintkey_stufftest_sv;
static int (*next_keyword_plugin)(pTHX_ char *, STRLEN, OP **);
/* low-level parser helpers */
@@ -150,6 +150,27 @@ static OP *THX_parse_keyword_calcrpn(pTHX)
}
#define parse_keyword_calcrpn() THX_parse_keyword_calcrpn(aTHX)
+static OP *THX_parse_keyword_stufftest(pTHX)
+{
+ I32 c;
+ bool do_stuff;
+ lex_read_space(0);
+ do_stuff = lex_peek_unichar(0) == '+';
+ if(do_stuff) {
+ lex_read_unichar(0);
+ lex_read_space(0);
+ }
+ c = lex_peek_unichar(0);
+ if(c == ';') {
+ lex_read_unichar(0);
+ } else if(c != /*{*/'}') {
+ croak("syntax error");
+ }
+ if(do_stuff) lex_stuff_pvn(" ", 1, 0);
+ return newOP(OP_NULL, 0);
+}
+#define parse_keyword_stufftest() THX_parse_keyword_stufftest(aTHX)
+
/* plugin glue */
static int THX_keyword_active(pTHX_ SV *hintkey_sv)
@@ -200,6 +221,10 @@ static int my_keyword_plugin(pTHX_
keyword_active(hintkey_calcrpn_sv)) {
*op_ptr = parse_keyword_calcrpn();
return KEYWORD_PLUGIN_STMT;
+ } else if(keyword_len == 9 && strnEQ(keyword_ptr, "stufftest", 9) &&
+ keyword_active(hintkey_stufftest_sv)) {
+ *op_ptr = parse_keyword_stufftest();
+ return KEYWORD_PLUGIN_STMT;
} else {
return next_keyword_plugin(aTHX_
keyword_ptr, keyword_len, op_ptr);
@@ -211,6 +236,8 @@ MODULE = XS::APItest::KeywordRPN PACKAGE = XS::APItest::KeywordRPN
BOOT:
hintkey_rpn_sv = newSVpvs_share("XS::APItest::KeywordRPN/rpn");
hintkey_calcrpn_sv = newSVpvs_share("XS::APItest::KeywordRPN/calcrpn");
+ hintkey_stufftest_sv =
+ newSVpvs_share("XS::APItest::KeywordRPN/stufftest");
next_keyword_plugin = PL_keyword_plugin;
PL_keyword_plugin = my_keyword_plugin;
@@ -225,6 +252,9 @@ PPCODE:
keyword_enable(hintkey_rpn_sv);
} else if(sv_is_string(item) && strEQ(SvPVX(item), "calcrpn")) {
keyword_enable(hintkey_calcrpn_sv);
+ } else if(sv_is_string(item) &&
+ strEQ(SvPVX(item), "stufftest")) {
+ keyword_enable(hintkey_stufftest_sv);
} else {
croak("\"%s\" is not exported by the %s module",
SvPV_nolen(item), SvPV_nolen(ST(0)));
@@ -242,6 +272,9 @@ PPCODE:
keyword_disable(hintkey_rpn_sv);
} else if(sv_is_string(item) && strEQ(SvPVX(item), "calcrpn")) {
keyword_disable(hintkey_calcrpn_sv);
+ } else if(sv_is_string(item) &&
+ strEQ(SvPVX(item), "stufftest")) {
+ keyword_disable(hintkey_stufftest_sv);
} else {
croak("\"%s\" is not exported by the %s module",
SvPV_nolen(item), SvPV_nolen(ST(0)));
diff --git a/ext/XS-APItest-KeywordRPN/t/stuff_svcur_bug.t b/ext/XS-APItest-KeywordRPN/t/stuff_svcur_bug.t
new file mode 100644
index 0000000..4fd6e11
--- /dev/null
+++ b/ext/XS-APItest-KeywordRPN/t/stuff_svcur_bug.t
@@ -0,0 +1,12 @@
+use warnings;
+use strict;
+
+use Test::More tests => 1;
+ok 1;
+
+use XS::APItest::KeywordRPN qw(stufftest);
+
+# In the buggy case, a syntax error occurs at EOF.
+# Adding a semicolon, any following statements, or anything else
+# causes the bug not to show itself.
+stufftest+;()
diff --git a/toke.c b/toke.c
index a7a71a4..7b057e7 100644
--- a/toke.c
+++ b/toke.c
@@ -956,6 +956,8 @@ Perl_lex_stuff_pvn(pTHX_ char *pv, STRLEN len, U32 flags)
lex_grow_linestr(SvCUR(PL_parser->linestr)+1+len+highhalf);
bufptr = PL_parser->bufptr;
Move(bufptr, bufptr+len+highhalf, PL_parser->bufend+1-bufptr, char);
+ SvCUR_set(PL_parser->linestr,
+ SvCUR(PL_parser->linestr) + len+highhalf);
PL_parser->bufend += len+highhalf;
for (p = pv; p != e; p++) {
U8 c = (U8)*p;
@@ -994,6 +996,8 @@ Perl_lex_stuff_pvn(pTHX_ char *pv, STRLEN len, U32 flags)
lex_grow_linestr(SvCUR(PL_parser->linestr)+1+len-highhalf);
bufptr = PL_parser->bufptr;
Move(bufptr, bufptr+len-highhalf, PL_parser->bufend+1-bufptr, char);
+ SvCUR_set(PL_parser->linestr,
+ SvCUR(PL_parser->linestr) + len-highhalf);
PL_parser->bufend += len-highhalf;
for (p = pv; p != e; p++) {
U8 c = (U8)*p;
@@ -1009,6 +1013,7 @@ Perl_lex_stuff_pvn(pTHX_ char *pv, STRLEN len, U32 flags)
lex_grow_linestr(SvCUR(PL_parser->linestr)+1+len);
bufptr = PL_parser->bufptr;
Move(bufptr, bufptr+len, PL_parser->bufend+1-bufptr, char);
+ SvCUR_set(PL_parser->linestr, SvCUR(PL_parser->linestr) + len);
PL_parser->bufend += len;
Copy(pv, bufptr, len, char);
} |
From @obraI'd prefer this that this go in for 5.12.1, not 5.12.0 |
|
The RT System itself - Status changed from 'new' to 'open' |
From zefram@fysh.orgJesse Vincent wrote:
Would someone please commit it now? -zefram |
|
@rgs - Status changed from 'open' to 'resolved' |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Migrated from rt.perl.org#74006 (status was 'resolved')
Searchable as RT74006$
The text was updated successfully, but these errors were encountered: