Changes from Jonathan Riehl to allow his pgen extension (PEP 269) to
work. This includes some more code that used to be part of pgen in
the main parser; I'm okay with that. I'll see if the Windows build
needs work next.
diff --git a/Include/parsetok.h b/Include/parsetok.h
index 99a79b7..b788566 100644
--- a/Include/parsetok.h
+++ b/Include/parsetok.h
@@ -38,6 +38,10 @@
const char *,
grammar *, int,
perrdetail *, int);
+
+/* Note that he following function is defined in pythonrun.c not parsetok.c. */
+PyAPI_FUNC(void) PyParser_SetError(perrdetail *);
+
#ifdef __cplusplus
}
#endif
diff --git a/Include/pgen.h b/Include/pgen.h
new file mode 100644
index 0000000..8a325ed
--- /dev/null
+++ b/Include/pgen.h
@@ -0,0 +1,18 @@
+#ifndef Py_PGEN_H
+#define Py_PGEN_H
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+
+/* Parser generator interface */
+
+extern grammar *meta_grammar(void);
+
+struct _node;
+extern grammar *pgen(struct _node *);
+
+#ifdef __cplusplus
+}
+#endif
+#endif /* !Py_PGEN_H */
diff --git a/Makefile.pre.in b/Makefile.pre.in
index 5313030..d461346 100644
--- a/Makefile.pre.in
+++ b/Makefile.pre.in
@@ -194,7 +194,10 @@
Parser/parser.o \
Parser/parsetok.o \
Parser/bitset.o \
- Parser/metagrammar.o
+ Parser/metagrammar.o \
+ Parser/firstsets.o \
+ Parser/grammar.o \
+ Parser/pgen.o
PARSER_OBJS= $(POBJS) Parser/myreadline.o Parser/tokenizer.o
@@ -202,9 +205,6 @@
Objects/obmalloc.o \
Python/mysnprintf.o \
Parser/tokenizer_pgen.o \
- Parser/firstsets.o \
- Parser/grammar.o \
- Parser/pgen.o \
Parser/printgrammar.o \
Parser/pgenmain.o
diff --git a/Parser/grammar.c b/Parser/grammar.c
index 2cc15b58..1f01562 100644
--- a/Parser/grammar.c
+++ b/Parser/grammar.c
@@ -38,7 +38,7 @@
Py_FatalError("no mem to resize dfa in adddfa");
d = &g->g_dfa[g->g_ndfas++];
d->d_type = type;
- d->d_name = name;
+ d->d_name = strdup(name);
d->d_nstates = 0;
d->d_state = NULL;
d->d_initial = -1;
@@ -98,7 +98,10 @@
Py_FatalError("no mem to resize labellist in addlabel");
lb = &ll->ll_label[ll->ll_nlabels++];
lb->lb_type = type;
- lb->lb_str = str; /* XXX strdup(str) ??? */
+ lb->lb_str = strdup(str);
+ if (Py_DebugFlag)
+ printf("Label @ %08x, %d: %s\n", (unsigned)ll, ll->ll_nlabels,
+ PyGrammar_LabelRepr(lb));
return lb - ll->ll_label;
}
@@ -152,6 +155,7 @@
lb->lb_str,
g->g_dfa[i].d_type);
lb->lb_type = g->g_dfa[i].d_type;
+ free(lb->lb_str);
lb->lb_str = NULL;
return;
}
@@ -162,6 +166,7 @@
printf("Label %s is terminal %d.\n",
lb->lb_str, i);
lb->lb_type = i;
+ free(lb->lb_str);
lb->lb_str = NULL;
return;
}
@@ -173,18 +178,29 @@
if (lb->lb_type == STRING) {
if (isalpha((int)(lb->lb_str[1])) || lb->lb_str[1] == '_') {
char *p;
+ char *src;
+ char *dest;
+ size_t name_len;
if (Py_DebugFlag)
printf("Label %s is a keyword\n", lb->lb_str);
lb->lb_type = NAME;
- lb->lb_str++;
- p = strchr(lb->lb_str, '\'');
+ src = lb->lb_str + 1;
+ p = strchr(src, '\'');
if (p)
- *p = '\0';
+ name_len = p - src;
+ else
+ name_len = strlen(src);
+ dest = malloc(name_len + 1);
+ strncpy(dest, src, name_len);
+ dest[name_len] = '\0';
+ free(lb->lb_str);
+ lb->lb_str = dest;
}
else if (lb->lb_str[2] == lb->lb_str[0]) {
int type = (int) PyToken_OneChar(lb->lb_str[1]);
if (type != OP) {
lb->lb_type = type;
+ free(lb->lb_str);
lb->lb_str = NULL;
}
else
@@ -196,6 +212,7 @@
lb->lb_str[2]);
if (type != OP) {
lb->lb_type = type;
+ free(lb->lb_str);
lb->lb_str = NULL;
}
else
@@ -208,6 +225,7 @@
lb->lb_str[3]);
if (type != OP) {
lb->lb_type = type;
+ free(lb->lb_str);
lb->lb_str = NULL;
}
else
diff --git a/Parser/metagrammar.c b/Parser/metagrammar.c
index ae67eb1..b61bc6d 100644
--- a/Parser/metagrammar.c
+++ b/Parser/metagrammar.c
@@ -151,3 +151,9 @@
{
return &_PyParser_Grammar;
}
+
+grammar *
+Py_meta_grammar(void)
+{
+ return meta_grammar();
+}
diff --git a/Parser/pgen.c b/Parser/pgen.c
index f3fdb46..453deb4 100644
--- a/Parser/pgen.c
+++ b/Parser/pgen.c
@@ -669,6 +669,11 @@
return g;
}
+grammar *
+Py_pgen(node *n)
+{
+ return pgen(n);
+}
/*