diff options
author | David Soria Parra <dsp@php.net> | 2007-10-03 10:44:32 +0000 |
---|---|---|
committer | David Soria Parra <dsp@php.net> | 2007-10-03 10:44:32 +0000 |
commit | d4e8fa62c6db480f5926661302b71e7f0a8dca3c (patch) | |
tree | bcc4f44c3c4d6b08c732c1d2173325859c2b76d3 /ext | |
parent | eb0c56ada1e2d6f4155c32c23cea56502de7751e (diff) | |
download | php-git-d4e8fa62c6db480f5926661302b71e7f0a8dca3c.tar.gz |
MFH: Add an escape parameter to fgetcsv to satisfy rfc4180 and bug #40501.
[DOC] 5th parameter available since PHP 5.3+
Diffstat (limited to 'ext')
-rwxr-xr-x | ext/spl/spl_directory.c | 43 | ||||
-rwxr-xr-x | ext/spl/spl_directory.h | 1 | ||||
-rw-r--r-- | ext/standard/file.c | 32 | ||||
-rw-r--r-- | ext/standard/file.h | 2 | ||||
-rw-r--r-- | ext/standard/tests/file/bug40501.csv | 2 | ||||
-rw-r--r-- | ext/standard/tests/file/bug40501.phpt | 20 | ||||
-rw-r--r-- | ext/standard/tests/file/fgetcsv_error.phpt | 9 |
7 files changed, 82 insertions, 27 deletions
diff --git a/ext/spl/spl_directory.c b/ext/spl/spl_directory.c index ccf69e2409..beef441c02 100755 --- a/ext/spl/spl_directory.c +++ b/ext/spl/spl_directory.c @@ -1589,7 +1589,7 @@ static int spl_filesystem_file_call(spl_filesystem_object *intern, zend_function spl_filesystem_file_call(intern, func_ptr, pass_num_args, return_value, arg2 TSRMLS_CC); \ } -static int spl_filesystem_file_read_csv(spl_filesystem_object *intern, char delimiter, char enclosure, zval *return_value TSRMLS_DC) /* {{{ */ +static int spl_filesystem_file_read_csv(spl_filesystem_object *intern, char delimiter, char enclosure, char escape, zval *return_value TSRMLS_DC) /* {{{ */ { int ret = SUCCESS; @@ -1606,7 +1606,7 @@ static int spl_filesystem_file_read_csv(spl_filesystem_object *intern, char deli } ALLOC_INIT_ZVAL(intern->u.file.current_zval); - php_fgetcsv(intern->u.file.stream, delimiter, enclosure, buf_len, buf, intern->u.file.current_zval TSRMLS_CC); + php_fgetcsv(intern->u.file.stream, delimiter, enclosure, escape, buf_len, buf, intern->u.file.current_zval TSRMLS_CC); if (return_value) { if (Z_TYPE_P(return_value) != IS_NULL) { zval_dtor(return_value); @@ -1632,7 +1632,7 @@ static int spl_filesystem_file_read_line_ex(zval * this_ptr, spl_filesystem_obje return FAILURE; } if (intern->flags & SPL_FILE_OBJECT_READ_CSV) { - return spl_filesystem_file_read_csv(intern, intern->u.file.delimiter, intern->u.file.enclosure, NULL TSRMLS_CC); + return spl_filesystem_file_read_csv(intern, intern->u.file.delimiter, intern->u.file.enclosure, intern->u.file.escape, NULL TSRMLS_CC); } else { zend_call_method_with_0_params(&this_ptr, Z_OBJCE_P(getThis()), &intern->u.file.func_getCurr, "getCurrentLine", &retval); } @@ -1943,18 +1943,25 @@ SPL_METHOD(SplFileObject, func_name) \ } /* }}} */ -/* {{{ proto array SplFileObject::fgetcsv([string delimiter [, string enclosure]]) +/* {{{ proto array SplFileObject::fgetcsv([string delimiter [, string enclosure [, escape = '\\']]]) Return current line as csv */ SPL_METHOD(SplFileObject, fgetcsv) { spl_filesystem_object *intern = (spl_filesystem_object*)zend_object_store_get_object(getThis() TSRMLS_CC); - char delimiter = intern->u.file.delimiter, enclosure = intern->u.file.enclosure; - char *delim, *enclo; - int d_len, e_len; + char delimiter = intern->u.file.delimiter, enclosure = intern->u.file.enclosure, escape = intern->u.file.escape; + char *delim, *enclo, *esc; + int d_len, e_len, esc_len; - if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "|ss", &delim, &d_len, &enclo, &e_len) == SUCCESS) { + if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "|sss", &delim, &d_len, &enclo, &e_len, &esc, &esc_len) == SUCCESS) { switch(ZEND_NUM_ARGS()) { + case 3: + if (esc_len != 1) { + php_error_docref(NULL TSRMLS_CC, E_WARNING, "escape must be a character"); + RETURN_FALSE; + } + escape = esc[0]; + /* no break */ case 2: if (e_len != 1) { php_error_docref(NULL TSRMLS_CC, E_WARNING, "enclosure must be a character"); @@ -1972,23 +1979,30 @@ SPL_METHOD(SplFileObject, fgetcsv) case 0: break; } - spl_filesystem_file_read_csv(intern, delimiter, enclosure, return_value TSRMLS_CC); + spl_filesystem_file_read_csv(intern, delimiter, enclosure, escape, return_value TSRMLS_CC); } } /* }}} */ -/* {{{ proto void SplFileObject::setCsvControl([string delimiter = ',' [, string enclosure = '"']]) +/* {{{ proto void SplFileObject::setCsvControl([string delimiter = ',' [, string enclosure = '"' [, string escape = '\\']]]) Set the delimiter and enclosure character used in fgetcsv */ SPL_METHOD(SplFileObject, setCsvControl) { spl_filesystem_object *intern = (spl_filesystem_object*)zend_object_store_get_object(getThis() TSRMLS_CC); - char delimiter = ',', enclosure = '"'; - char *delim, *enclo; - int d_len, e_len; + char delimiter = ',', enclosure = '"', escape='\\'; + char *delim, *enclo, *esc; + int d_len, e_len, esc_len; - if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "|ss", &delim, &d_len, &enclo, &e_len) == SUCCESS) { + if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "|sss", &delim, &d_len, &enclo, &e_len, &esc, &esc_len) == SUCCESS) { switch(ZEND_NUM_ARGS()) { + case 3: + if (esc_len != 1) { + php_error_docref(NULL TSRMLS_CC, E_WARNING, "escape must be a character"); + RETURN_FALSE; + } + escape = esc[0]; + /* no break */ case 2: if (e_len != 1) { php_error_docref(NULL TSRMLS_CC, E_WARNING, "enclosure must be a character"); @@ -2008,6 +2022,7 @@ SPL_METHOD(SplFileObject, setCsvControl) } intern->u.file.delimiter = delimiter; intern->u.file.enclosure = enclosure; + intern->u.file.escape = escape; } } /* }}} */ diff --git a/ext/spl/spl_directory.h b/ext/spl/spl_directory.h index 172c400ce1..caf3a18167 100755 --- a/ext/spl/spl_directory.h +++ b/ext/spl/spl_directory.h @@ -84,6 +84,7 @@ struct _spl_filesystem_object { zend_function *func_getCurr; char delimiter; char enclosure; + char escape; } file; } u; }; diff --git a/ext/standard/file.c b/ext/standard/file.c index c54ce62313..ce4e854d7f 100644 --- a/ext/standard/file.c +++ b/ext/standard/file.c @@ -2013,12 +2013,14 @@ PHP_FUNCTION(fputcsv) } /* }}} */ -/* {{{ proto array fgetcsv(resource fp [,int length [, string delimiter [, string enclosure]]]) +/* {{{ proto array fgetcsv(resource fp [,int length [, string delimiter [, string enclosure [, string escape]]]]) Get line from file pointer and parse for CSV fields */ PHP_FUNCTION(fgetcsv) { char delimiter = ','; /* allow this to be set as parameter */ char enclosure = '"'; /* allow this to be set as parameter */ + char escape = '\\'; + /* first section exactly as php_fgetss */ long len = 0; @@ -2032,10 +2034,13 @@ PHP_FUNCTION(fgetcsv) int delimiter_str_len = 0; char *enclosure_str = NULL; int enclosure_str_len = 0; + char *escape_str = NULL; + int escape_str_len = 0; - if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "r|Zss", - &fd, &len_zv, &delimiter_str, &delimiter_str_len, - &enclosure_str, &enclosure_str_len) == FAILURE) { + if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "r|Zsss", + &fd, &len_zv, &delimiter_str, &delimiter_str_len, + &enclosure_str, &enclosure_str_len, + &escape_str, &escape_str_len) == FAILURE) { return; } @@ -2064,6 +2069,17 @@ PHP_FUNCTION(fgetcsv) enclosure = enclosure_str[0]; } + if (escape_str != NULL) { + if (escape_str_len < 1) { + php_error_docref(NULL TSRMLS_CC, E_WARNING, "escape must be character"); + RETURN_FALSE; + } else if (escape_str_len > 1) { + php_error_docref(NULL TSRMLS_CC, E_NOTICE, "escape must be a single character"); + } + + escape = escape_str[0]; + } + if (len_zv != NULL && Z_TYPE_PP(len_zv) != IS_NULL) { convert_to_long_ex(len_zv); len = Z_LVAL_PP(len_zv); @@ -2092,14 +2108,13 @@ PHP_FUNCTION(fgetcsv) } } - php_fgetcsv(stream, delimiter, enclosure, buf_len, buf, return_value TSRMLS_CC); + php_fgetcsv(stream, delimiter, enclosure, escape, buf_len, buf, return_value TSRMLS_CC); } /* }}} */ -PHPAPI void php_fgetcsv(php_stream *stream, char delimiter, char enclosure, size_t buf_len, char *buf, zval *return_value TSRMLS_DC) /* {{{ */ +PHPAPI void php_fgetcsv(php_stream *stream, char delimiter, char enclosure, char escape_char, size_t buf_len, char *buf, zval *return_value TSRMLS_DC) /* {{{ */ { char *temp, *tptr, *bptr, *line_end, *limit; - const char escape_char = '\\'; size_t temp_len, line_end_len; int inc_len; @@ -2250,7 +2265,8 @@ PHPAPI void php_fgetcsv(php_stream *stream, char delimiter, char enclosure, size state = 0; break; default: - if (*bptr == escape_char) { + if ((escape_char == enclosure && *bptr == escape_char && *(bptr+1) == escape_char) + || (escape_char != enclosure && *bptr == escape_char)) { state = 1; } else if (*bptr == enclosure) { state = 2; diff --git a/ext/standard/file.h b/ext/standard/file.h index 2b6fe2f2d1..38cdca8a7f 100644 --- a/ext/standard/file.h +++ b/ext/standard/file.h @@ -78,7 +78,7 @@ PHPAPI int php_copy_file(char *src, char *dest TSRMLS_DC); PHPAPI int php_copy_file_ex(char *src, char *dest, int src_chk TSRMLS_DC); PHPAPI int php_mkdir_ex(char *dir, long mode, int options TSRMLS_DC); PHPAPI int php_mkdir(char *dir, long mode TSRMLS_DC); -PHPAPI void php_fgetcsv(php_stream *stream, char delimiter, char enclosure, size_t buf_len, char *buf, zval *return_value TSRMLS_DC); +PHPAPI void php_fgetcsv(php_stream *stream, char delimiter, char enclosure, char escape_char, size_t buf_len, char *buf, zval *return_value TSRMLS_DC); #define META_DEF_BUFSIZE 8192 diff --git a/ext/standard/tests/file/bug40501.csv b/ext/standard/tests/file/bug40501.csv new file mode 100644 index 0000000000..c786ed98e8 --- /dev/null +++ b/ext/standard/tests/file/bug40501.csv @@ -0,0 +1,2 @@ +"this element contains the delimiter, and ends with an odd number of +backslashes (ex: 1)\",and it isn't the last element$
\ No newline at end of file diff --git a/ext/standard/tests/file/bug40501.phpt b/ext/standard/tests/file/bug40501.phpt new file mode 100644 index 0000000000..110533fb82 --- /dev/null +++ b/ext/standard/tests/file/bug40501.phpt @@ -0,0 +1,20 @@ +--TEST-- +Bug #40501 (fgetcsv() can't handle trailing odd number of backslashes) +--FILE-- +<?php +$file = dirname(__FILE__).'/bug40501.csv'; + +$h = fopen($file, 'r'); +$data = fgetcsv($h, NULL, ',', '"', '"'); +fclose($h); + +var_dump($data); +?> +--EXPECT-- +array(2) { + [0]=> + string(88) "this element contains the delimiter, and ends with an odd number of +backslashes (ex: 1)\" + [1]=> + string(30) "and it isn't the last element$" +}
\ No newline at end of file diff --git a/ext/standard/tests/file/fgetcsv_error.phpt b/ext/standard/tests/file/fgetcsv_error.phpt index c3ace9ab87..e2db9f8730 100644 --- a/ext/standard/tests/file/fgetcsv_error.phpt +++ b/ext/standard/tests/file/fgetcsv_error.phpt @@ -3,7 +3,7 @@ Test fgetcsv() function : error conditions --FILE-- <?php /* - Prototype: array fgetcsv ( resource $handle [, int $length [, string $delimiter [, string $enclosure]]] ); + Prototype: array fgetcsv ( resource $handle [, int $length [, string $delimiter [, string $enclosure [, string $escape]]]] ); Description: Gets line from file pointer and parse for CSV fields */ @@ -18,7 +18,8 @@ $fp = fopen(__FILE__, "r"); $len = 1024; $delim = ";"; $enclosure ="\""; -var_dump( fgetcsv($fp, $len, $delim, $enclosure, $fp) ); +$escape = '"'; +var_dump( fgetcsv($fp, $len, $delim, $enclosure, $escape, $fp) ); fclose($fp); // test invalid arguments : non-resources @@ -35,7 +36,7 @@ $invalid_args = array ( for($loop_counter = 1; $loop_counter <= count($invalid_args); $loop_counter++) { echo "-- Iteration $loop_counter --\n"; var_dump( fgetcsv($invalid_args[$loop_counter - 1]) ); // with default args - var_dump( fgetcsv($invalid_args[$loop_counter - 1], $len, $delim, $enclosure) ); // all args specified + var_dump( fgetcsv($invalid_args[$loop_counter - 1], $len, $delim, $enclosure, $escape) ); // all args specified } echo "Done\n"; @@ -47,7 +48,7 @@ Warning: fgetcsv() expects at least 1 parameter, 0 given in %s on line %d NULL -- Testing fgetcsv() with more than expected number of arguments -- -Warning: fgetcsv() expects at most 4 parameters, 5 given in %s on line %d +Warning: fgetcsv() expects at most 5 parameters, 6 given in %s on line %d NULL -- Testing fgetcsv() with invalid arguments -- -- Iteration 1 -- |