Add str_getcsv() and fix a couple cases in recent fgetcsv() reimplementation

diff --git a/NEWS b/NEWS
index d31dab67743f3a3d4381353130ff89092c6677ea..d8fbce1721cbfb60cd597a34763482c7a9f53fba 100644 (file)
--- a/NEWS
+++ b/NEWS
@@ -64,6 +64,7 @@ PHP                                                                        NEWS
   . "context" and "binary_pipes" params in "other_options" arg. (Sara)
   . stream_resolve_include_path(). (Sara)
 - Added shm_has_var() function. (Mike)
+- Added str_getcsv() function. (Sara)
 
 - Fixed bug #39504 (xmlwriter_write_dtd_entity() creates Attlist tag, 
   not enity). (Hannes Magnusson)

diff --git a/ext/standard/basic_functions.c b/ext/standard/basic_functions.c
index 2891d0e457316d8036f19813df3949cbe962359f..4924f42d7ccdacd6850a9691c0f2a5e8ba3e03fc 100644 (file)
--- a/ext/standard/basic_functions.c
+++ b/ext/standard/basic_functions.c
@@ -2751,6 +2751,14 @@ ZEND_BEGIN_ARG_INFO_EX(arginfo_parse_str, 0, 0, 1)
        ZEND_ARG_INFO(1, result)
 ZEND_END_ARG_INFO()
 
+static
+ZEND_BEGIN_ARG_INFO_EX(arginfo_str_getcsv, 0, 0, 1)
+       ZEND_ARG_INFO(0, string)
+       ZEND_ARG_INFO(0, delimiter)
+       ZEND_ARG_INFO(0, enclosure)
+       ZEND_ARG_INFO(0, escape)
+ZEND_END_ARG_INFO()
+
 static
 ZEND_BEGIN_ARG_INFO(arginfo_str_repeat, 0)
        ZEND_ARG_INFO(0, input)
@@ -3230,6 +3238,7 @@ zend_function_entry basic_functions[] = {
        PHP_FE(chr,                                                                                                                             arginfo_chr)
        PHP_FE(ord,                                                                                                                             arginfo_ord)
        PHP_FE(parse_str,                                                                                                               arginfo_parse_str)
+       PHP_FE(str_getcsv,                                                                                                              arginfo_str_getcsv)
        PHP_FE(str_pad,                                                                                                                 arginfo_str_pad)
        PHP_FALIAS(chop,                                rtrim,                                                                  arginfo_rtrim)
        PHP_FALIAS(strchr,                              strstr,                                                                 arginfo_strstr)

diff --git a/ext/standard/file.c b/ext/standard/file.c
index eff25d0e756206559d008db168aa82a2261d2055..c035722609c534e3dd2d0f9dcaf14c016a818e2a 100644 (file)
--- a/ext/standard/file.c
+++ b/ext/standard/file.c
@@ -2423,11 +2423,21 @@ post_enc:
 
                                /* Simple character */
                                p++;
+
+                               if (p == e) {
+                                       add_next_index_stringl(return_value, field_start, p - field_start, 1);
+                                       /* Reset scanner even though we're dying */
+                                       state = PHP_FGETCSV_READY;
+                                       field_start = field_end = NULL;
+                                       p += delimiter_len;
+                               }
                                break;
                }
        }
 
-       efree(buffer);
+       if (stream) {
+               efree(buffer);
+       }
 }
 /* }}} */
 
@@ -2617,11 +2627,21 @@ post_enc:
 
                                /* Simple character */
                                p++;
+
+                               if (p == e) {
+                                       add_next_index_unicodel(return_value, field_start, p - field_start, 1);
+                                       /* Reset scanner even though we're dying */
+                                       state = PHP_FGETCSV_READY;
+                                       field_start = field_end = NULL;
+                                       p += delimiter_len;
+                               }
                                break;
                }
        }
 
-       efree(buffer);
+       if (stream) {
+               efree(buffer);
+       }
 }
 /* }}} */
 

diff --git a/ext/standard/php_string.h b/ext/standard/php_string.h
index c33fd679c494ae6c9b1e04f54c5611d6ffa9c196..909c3d29d11a88a9d30f4d5183666d89321757f9 100644 (file)
--- a/ext/standard/php_string.h
+++ b/ext/standard/php_string.h
@@ -76,6 +76,7 @@ PHP_FUNCTION(nl_langinfo);
 PHP_FUNCTION(stristr);
 PHP_FUNCTION(chunk_split);
 PHP_FUNCTION(parse_str);
+PHP_FUNCTION(str_getcsv);
 PHP_FUNCTION(bin2hex);
 PHP_FUNCTION(similar_text);
 PHP_FUNCTION(strip_tags);

diff --git a/ext/standard/string.c b/ext/standard/string.c
index ab610e0a79d15bc265d1e34a85d63da58be8e1f8..4adfb4cff82064bc29adc9d08745eb4e574a9681 100644 (file)
--- a/ext/standard/string.c
+++ b/ext/standard/string.c
@@ -56,6 +56,9 @@
 #include "unicode/uchar.h"
 #include "unicode/ubrk.h"
 
+/* For str_getcsv() support */
+#include "ext/standard/file.h"
+
 #define STR_PAD_LEFT                   0
 #define STR_PAD_RIGHT                  1
 #define STR_PAD_BOTH                   2
@@ -6734,6 +6737,46 @@ reg_char:
 }
 /* }}} */
 
+/* {{{ proto array str_getcsv(string input[, string delimiter[, string enclosure[, string escape]]]) U
+Parse a CSV string into an array */
+PHP_FUNCTION(str_getcsv)
+{
+       zend_uchar str_type, delim_type = IS_STRING, enc_type = IS_STRING, esc_type = IS_STRING;
+       char *str, *delim = ",", *enc = "\"", *esc = "\\";
+       int str_len, delim_len = 1, enc_len = 1, esc_len = 1;
+
+       if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "T|TTT",
+                       &str,   &str_len,   &str_type,
+                       &delim, &delim_len, &delim_type,
+                       &enc,   &enc_len,   &enc_type,
+                       &esc,   &esc_len,   &esc_type) == FAILURE) {
+               return;
+       }
+
+       if (str_type == IS_UNICODE) {
+               UChar udelim = ',', uenc = '"', uesc = '\\';
+
+               /* Non-passed params would need to be upconverted, but we can cheat with some local declarations */
+               if (delim_type == IS_STRING) {
+                       delim = (char*)&udelim;
+                       delim_len = 1;
+               }
+               if (enc_type == IS_STRING) {
+                       enc = (char*)&uenc;
+                       enc_len = 1;
+               }
+               if (esc_type == IS_STRING) {
+                       esc = (char*)&uesc;
+                       esc_len = 1;
+               }
+
+               php_u_fgetcsv(NULL, (UChar*)delim, delim_len, (UChar*)enc, enc_len, (UChar*)esc, esc_len, (UChar*)str, str_len, return_value TSRMLS_CC);
+       } else {
+               php_fgetcsv_ex(NULL, delim, delim_len, enc, enc_len, esc, esc_len, str, str_len, return_value TSRMLS_CC);
+       }
+}
+/* }}} */
+
 /* {{{ proto string str_repeat(string input, int mult) U
    Returns the input string repeat mult times */
 PHP_FUNCTION(str_repeat)