aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
Diffstat
-rw-r--r--README2
-rw-r--r--libparser-generate.c2
-rw-r--r--libparser.72
-rw-r--r--libparser.c55
-rw-r--r--print-syntax.c44
5 files changed, 101 insertions, 4 deletions
diff --git a/README b/README
index 23d231b..b5b833e 100644
--- a/README
+++ b/README
@@ -52,7 +52,7 @@ EXTENDED DESCRIPTION
(* STRINGS *)
- _escape_simple = "\\" | "\"" | "'" | "a" | "b" | "f" | "n" | "r" | "v";
+ _escape_simple = "\\" | "\"" | "'" | "a" | "b" | "f" | "n" | "r" | "t" | "v";
_escape_hex = ("x" | "X"), _xdigit, _xdigit;
_escape_octal = _octal, {_octal}; (* May not exceed 255 in base 10 *)
_escape = _escape_simple | _escape_hex | _escape_octal | -;
diff --git a/libparser-generate.c b/libparser-generate.c
index 0a91835..47d2ebd 100644
--- a/libparser-generate.c
+++ b/libparser-generate.c
@@ -902,6 +902,8 @@ again:
tokens[i]->s[1] = '\n';
} else if (tokens[i]->s[2] == 'r') {
tokens[i]->s[1] = '\r';
+ } else if (tokens[i]->s[2] == 't') {
+ tokens[i]->s[1] = '\t';
} else if (tokens[i]->s[2] == 'v') {
tokens[i]->s[1] = '\v';
} else if (tokens[i]->s[2] == 'x' && isxdigit(tokens[i]->s[3]) && isxdigit(tokens[i]->s[4])) {
diff --git a/libparser.7 b/libparser.7
index c50b5c9..174bac9 100644
--- a/libparser.7
+++ b/libparser.7
@@ -64,7 +64,7 @@ identifier = _identifier_head, {_identifier_tail};
(* STRINGS *)
-_escape_simple = \(dq\e\e\(dq | \(dq\e\(dq\(dq | \(dq'\(dq | \(dqa\(dq | \(dqb\(dq | \(dqf\(dq | \(dqn\(dq | \(dqr\(dq | \(dqv\(dq;
+_escape_simple = \(dq\e\e\(dq | \(dq\e\(dq\(dq | \(dq'\(dq | \(dqa\(dq | \(dqb\(dq | \(dqf\(dq | \(dqn\(dq | \(dqr\(dq | \(dqt\(dq | \(dqv\(dq;
_escape_hex = (\(dqx\(dq | \(dqX\(dq), _xdigit, _xdigit;
_escape_octal = _octal, {_octal}; (* May not exceed 255 in base 10 *)
_escape = _escape_simple | _escape_hex | _escape_octal | -;
diff --git a/libparser.c b/libparser.c
index 5ecb7ee..2a925b7 100644
--- a/libparser.c
+++ b/libparser.c
@@ -363,6 +363,54 @@ print_state(struct context *ctx)
static int
+print_text(const char *text, size_t textlen)
+{
+ size_t off = 0, n = 0;
+ int len, ret = 0;
+
+ while (off + n < textlen) {
+ if (!text[off + n])
+ break;
+ if (text[off + n] < ' ' || text[off + n] >= 0x7F ||
+ text[off + n] == '"' || text[off + n] == '\\') {
+ fprintf(stderr, "%.*s%n", (int)n, &text[off], &len);
+ ret += len;
+ off += n;
+ n = 0;
+ switch (text[off]) {
+ case '\\': fprintf(stderr, "\\\\%n", &len); break;
+ case '\"': fprintf(stderr, "\\\"%n", &len); break;
+ case '\a': fprintf(stderr, "\\a%n", &len); break;
+ case '\b': fprintf(stderr, "\\b%n", &len); break;
+ case '\f': fprintf(stderr, "\\f%n", &len); break;
+ case '\n': fprintf(stderr, "\\n%n", &len); break;
+ case '\r': fprintf(stderr, "\\r%n", &len); break;
+ case '\t': fprintf(stderr, "\\t%n", &len); break;
+ case '\v': fprintf(stderr, "\\v%n", &len); break;
+ default:
+ fprintf(stderr, "\\x%02X%n", +(unsigned char)text[off], &len);
+ break;
+ }
+ off++;
+ } else if (n == 4096U) {
+ fprintf(stderr, "%.*s%n", (int)n, &text[off], &len);
+ ret += len;
+ off += n;
+ n = 0;
+ } else {
+ n++;
+ }
+ }
+ if (n) {
+ fprintf(stderr, "%.*s%n", (int)n, &text[off], &len);
+ ret += len;
+ }
+
+ return ret;
+}
+
+
+static int
print_sentence(const union libparser_sentence *sentence, int indent)
{
unsigned char low, high;
@@ -431,7 +479,9 @@ print_sentence(const union libparser_sentence *sentence, int indent)
break;
case LIBPARSER_SENTENCE_TYPE_STRING:
- fprintf(stderr, "\"%.*s\"%n", (int)sentence->string.length, sentence->string.string, &len);
+ fprintf(stderr, "\"%n", &len);
+ indent += len + print_text(sentence->string.string, sentence->string.length);
+ fprintf(stderr, "\"%n", &len);
indent += len;
break;
@@ -1200,6 +1250,9 @@ libparser_parse_file(const struct libparser_rule *const rules[], const char *dat
#if PRINT_ACTIONS
print_grammar(rules);
+ fprintf(stderr, "Input text: \"");
+ print_text(ctx.data, ctx.length);
+ fprintf(stderr, "\"\n");
#endif
/* TODO guard against left-side recursion */
diff --git a/print-syntax.c b/print-syntax.c
index 07b3858..6a00fb8 100644
--- a/print-syntax.c
+++ b/print-syntax.c
@@ -14,6 +14,7 @@
static int
print_sentence(const union libparser_sentence *sentence, int indent)
{
+ size_t off, n;
int len;
switch (sentence->type) {
@@ -79,7 +80,48 @@ print_sentence(const union libparser_sentence *sentence, int indent)
break;
case LIBPARSER_SENTENCE_TYPE_STRING:
- printf("\"%.*s\"%n", (int)sentence->string.length, sentence->string.string, &len);
+ printf("\"%n", &len);
+ indent += len;
+ off = 0;
+ n = 0;
+ while (off + n < sentence->string.length) {
+ if (sentence->string.string[off + n] < ' ' ||
+ sentence->string.string[off + n] >= 0x7F ||
+ sentence->string.string[off + n] == '"' ||
+ sentence->string.string[off + n] == '\\') {
+ printf("%.*s%n", (int)n, &sentence->string.string[off], &len);
+ indent += len;
+ off += n;
+ n = 0;
+ switch (sentence->string.string[off]) {
+ case '\\': printf("\\\\%n", &len); break;
+ case '\"': printf("\\\"%n", &len); break;
+ case '\a': printf("\\a%n", &len); break;
+ case '\b': printf("\\b%n", &len); break;
+ case '\f': printf("\\f%n", &len); break;
+ case '\n': printf("\\n%n", &len); break;
+ case '\r': printf("\\r%n", &len); break;
+ case '\t': printf("\\t%n", &len); break;
+ case '\v': printf("\\v%n", &len); break;
+ default:
+ printf("\\x%02X%n", +(unsigned char)sentence->string.string[off], &len);
+ break;
+ }
+ off++;
+ } else if (n == 4096U) {
+ printf("%.*s%n", (int)n, &sentence->string.string[off], &len);
+ indent += len;
+ off += n;
+ n = 0;
+ } else {
+ n++;
+ }
+ }
+ if (n) {
+ printf("%.*s%n", (int)n, &sentence->string.string[off], &len);
+ indent += len;
+ }
+ printf("\"%n", &len);
indent += len;
break;
generated by cgit v1.2.3-70-g09d2 (git 2.51.0) at 2026-02-27 20:12:05 +0000