Improve REPL detecting when input needs to continue.

Full CPython compatibility with this requires actually parsing the
input so far collected, and if it fails parsing due to lack of tokens,
then continue collecting input.  It's not worth doing it this way.  Not
having compatibility at this level does not hurt the goals of Micro
Python.

py/repl.c

@@ -14,43 +14,66 @@ bool str_startswith_word(const char *str, const char *head) {
     return head[i] == '\0' && (str[i] == '\0' || !unichar_isalpha(str[i]));
 }
 
-bool mp_repl_is_compound_stmt(const char *line) {
-    // compound if line starts with a certain keyword
-    if (
-           str_startswith_word(line, "if")
-        || str_startswith_word(line, "while")
-        || str_startswith_word(line, "for")
-        || str_startswith_word(line, "try")
-        || str_startswith_word(line, "with")
-        || str_startswith_word(line, "def")
-        || str_startswith_word(line, "class")
-        || str_startswith_word(line, "@")
-       ) {
-        return true;
+bool mp_repl_continue_with_input(const char *input) {
+    // check for blank input
+    if (input[0] == '\0') {
+        return false;
     }
 
-    // also "compound" if unmatched open bracket or triple quote
+    // check if input starts with a certain keyword
+    bool starts_with_compound_keyword =
+           input[0] == '@'
+        || str_startswith_word(input, "if")
+        || str_startswith_word(input, "while")
+        || str_startswith_word(input, "for")
+        || str_startswith_word(input, "try")
+        || str_startswith_word(input, "with")
+        || str_startswith_word(input, "def")
+        || str_startswith_word(input, "class")
+        ;
+
+    // check for unmatched open bracket or triple quote
+    // TODO don't look at triple quotes inside single quotes
     int n_paren = 0;
     int n_brack = 0;
     int n_brace = 0;
     int in_triple_quote = 0;
-    for (const char *l = line; *l; l++) {
-        switch (*l) {
+    const char *i;
+    for (i = input; *i; i++) {
+        switch (*i) {
             case '(': n_paren += 1; break;
             case ')': n_paren -= 1; break;
             case '[': n_brack += 1; break;
             case ']': n_brack -= 1; break;
             case '{': n_brace += 1; break;
             case '}': n_brace -= 1; break;
+            case '\'':
+                if (in_triple_quote != '"' && i[1] == '\'' && i[2] == '\'') {
+                    i += 2;
+                    in_triple_quote = '\'' - in_triple_quote;
+                }
+                break;
             case '"':
-                if (l[1] == '"' && l[2] == '"') {
-                    l += 2;
-                    in_triple_quote = 1 - in_triple_quote;
+                if (in_triple_quote != '\'' && i[1] == '"' && i[2] == '"') {
+                    i += 2;
+                    in_triple_quote = '"' - in_triple_quote;
                 }
                 break;
         }
     }
-    return n_paren > 0 || n_brack > 0 || n_brace > 0 || in_triple_quote != 0;
+
+    // continue if unmatched brackets or quotes
+    if (n_paren > 0 || n_brack > 0 || n_brace > 0 || in_triple_quote != 0) {
+        return true;
+    }
+
+    // continue if compound keyword and last line was not empty
+    if (starts_with_compound_keyword && i[-1] != '\n') {
+        return true;
+    }
+
+    // otherwise, don't continue
+    return false;
 }
 
 #endif // MICROPY_ENABLE_REPL_HELPERS

py/repl.h

 #if MICROPY_ENABLE_REPL_HELPERS
-bool mp_repl_is_compound_stmt(const char *line);
+bool mp_repl_continue_with_input(const char *input);
 #endif

stm/pyexec.c

@@ -283,17 +283,14 @@ void pyexec_repl(void) {
             continue;
         }
 
-        if (mp_repl_is_compound_stmt(vstr_str(&line))) {
-            for (;;) {
+        while (mp_repl_continue_with_input(vstr_str(&line))) {
             vstr_add_char(&line, '\n');
-                int len = vstr_len(&line);
             int ret = readline(&line, "... ");
-                if (ret == VCP_CHAR_CTRL_D || vstr_len(&line) == len) {
-                    // done entering compound statement
+            if (ret == VCP_CHAR_CTRL_D) {
+                // stop entering compound statement
                 break;
             }
         }
-        }
 
         mp_lexer_t *lex = mp_lexer_new_from_str_len(MP_QSTR__lt_stdin_gt_, vstr_str(&line), vstr_len(&line), 0);
         parse_compile_execute(lex, MP_PARSE_SINGLE_INPUT, true);

stmhal/pyexec.c

@@ -204,17 +204,14 @@ friendly_repl_reset:
             continue;
         }
 
-        if (mp_repl_is_compound_stmt(vstr_str(&line))) {
-            for (;;) {
+        while (mp_repl_continue_with_input(vstr_str(&line))) {
             vstr_add_char(&line, '\n');
-                int len = vstr_len(&line);
             int ret = readline(&line, "... ");
-                if (ret == VCP_CHAR_CTRL_D || vstr_len(&line) == len) {
-                    // done entering compound statement
+            if (ret == VCP_CHAR_CTRL_D) {
+                // stop entering compound statement
                 break;
             }
         }
-        }
 
         mp_lexer_t *lex = mp_lexer_new_from_str_len(MP_QSTR__lt_stdin_gt_, vstr_str(&line), vstr_len(&line), 0);
         parse_compile_execute(lex, MP_PARSE_SINGLE_INPUT, true);

teensy/main.c

@@ -399,17 +399,14 @@ void do_repl(void) {
             continue;
         }
 
-        if (mp_repl_is_compound_stmt(vstr_str(&line))) {
-            for (;;) {
+        while (mp_repl_continue_with_input(vstr_str(&line))) {
             vstr_add_char(&line, '\n');
-                int len = vstr_len(&line);
             int ret = readline(&line, "... ");
-                if (ret == 0 || vstr_len(&line) == len) {
-                    // done entering compound statement
+            if (ret == 0) {
+                // stop entering compound statement
                 break;
             }
         }
-        }
 
         mp_lexer_t *lex = mp_lexer_new_from_str_len(MP_QSTR__lt_stdin_gt_, vstr_str(&line), vstr_len(&line), 0);
         mp_parse_error_kind_t parse_error_kind;

unix/main.c

@@ -146,10 +146,9 @@ STATIC void do_repl(void) {
             // EOF
             return;
         }
-        if (mp_repl_is_compound_stmt(line)) {
-            for (;;) {
+        while (mp_repl_continue_with_input(line)) {
             char *line2 = prompt("... ");
-                if (line2 == NULL || strlen(line2) == 0) {
+            if (line2 == NULL) {
                 break;
             }
             char *line3 = strjoin(line, '\n', line2);
@@ -157,7 +156,6 @@ STATIC void do_repl(void) {
             free(line2);
             line = line3;
         }
-        }
 
         mp_lexer_t *lex = mp_lexer_new_from_str_len(MP_QSTR__lt_stdin_gt_, line, strlen(line), false);
         execute_from_lexer(lex, MP_PARSE_SINGLE_INPUT, true);

windows/main.c

@@ -126,10 +126,9 @@ static void do_repl(void) {
             // EOF
             return;
         }
-        if (mp_repl_is_compound_stmt(line)) {
-            for (;;) {
+        while (mp_repl_continue_with_input(line)) {
             char *line2 = prompt("... ");
-                if (line2 == NULL || strlen(line2) == 0) {
+            if (line2 == NULL) {
                 break;
             }
             char *line3 = str_join(line, '\n', line2);
@@ -137,7 +136,6 @@ static void do_repl(void) {
             free(line2);
             line = line3;
         }
-        }
 
         mp_lexer_t *lex = mp_lexer_new_from_str_len(MP_QSTR__lt_stdin_gt_, line, strlen(line), false);
         execute_from_lexer(lex, MP_PARSE_SINGLE_INPUT, true);