On Tue, Apr 17, 2012 at 12:06 AM, Laruence <larue...@php.net> wrote:
> Hi:
>
> you can core dump the backtrace, then exam the related hash table to
> find out what's going wrong there.
So the problem doesn't seem to be with contains() at all. The problem is
actually coming from another method I implemented called
ArrayList::append(mixed $element). I figured it out by calling contains
first and I noticed it didn't cause a segfault. appends() seems rather
trivial but I guess I was wrong:
/* {{{ proto public boolean ArrayList::append(mixed $element)
Add an element to the end of the list */
ZEND_METHOD(arraymap_class, append) {
zval *object = getThis();
zval *element;
arraylist_object *intern;
if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "z", &element) ==
FAILURE) {
RETURN_FALSE;
}
intern = (arraylist_object *) zend_object_store_get_object(object
TSRMLS_CC);
if (add_next_index_zval(intern->elements, element) == SUCCESS) {
Z_ADDREF_P(element); /* increases ref count */
intern->size++;
RETURN_TRUE;
}
RETURN_FALSE;
}
This method returns true, as expected.
> and some maybe un-relevant issues:
> 1. plz put all var declarations at the begining of a block (C89)
> 2. do not use "this" keyword as variable name (it's C++ keyword)
>
fixed.
> thanks
>
> On Tue, Apr 17, 2012 at 2:12 PM, Yader Hernandez
> <yader.hernan...@gmail.com> wrote:
> > Hello,
> >
> > I'm trying to create an ArrayList as an extension. I'm currently
> > implementing ArrayList::contains(mixed $element) but I'm running into a
> > blocking issue.
> >
> > It's currently causing a segfault when you call contains. I've tried to
> > track down what I'm doing wrong, but I've had no luck with it.
> >
> > When running gdb this is what I get:
> >
> > Program received signal EXC_BAD_ACCESS, Could not access memory.
> > Reason: 13 at address: 0x00000000000000000x00000001005812ca in
> > _zend_is_inconsistent
> >
> > /* {{{ proto public boolean ArrayList::contain(mixed $element[, boolean
> > strict])
> > Returns true if this list contains the specified element. */
> > ZEND_METHOD(arraymap_class, contains) {
> > zval *element;
> > zend_bool strict = 0; /* strict comparison or not */
> >
> > if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "z|b", &element,
> > &strict) == FAILURE) {
> > RETURN_FALSE;
> > }
> >
> > zval *this = getThis();
> > arraylist_object *intern;
> >
> > intern = (arraylist_object *) zend_object_store_get_object(this
> > TSRMLS_CC);
> >
> > /* Don't use "ht" as a variable, it's already being used. All
> ZEND_METHOD
> > have "ht" defined. */
> > HashTable *hash_table = Z_ARRVAL_P(intern->elements);
> > HashPosition pos;
> > zval **current;
> > zval res; /* comparison result */
> >
> > int (*is_equal_func)(zval *, zval *, zval * TSRMLS_DC) =
> > is_equal_function;
> >
> > if (strict) {
> > is_equal_func = is_identical_function;
> > }
> >
> > zend_hash_internal_pointer_reset_ex(hash_table, &pos);
> >
> > while (zend_hash_get_current_data_ex(hash_table, (void **)¤t,
> &pos)
> > == SUCCESS) {
> > is_equal_func(&res, element, *current TSRMLS_CC);
> >
> > if (Z_LVAL(res)) {
> > RETURN_TRUE;
> > }
> >
> > zend_hash_move_forward_ex(hash_table, &pos);
> > }
> >
> > RETURN_FALSE;
> > }
> >
> > I can't seem to spot any errors with this implementation. The next thing
> I
> > thought that could be wrong is how I'm creating elements pointer, but
> that
> > seems correct too:
> >
> > static zend_object_value create_arraylist_object(zend_class_entry *ce
> > TSRMLS_DC) /* {{{ */
> > {
> > zend_object_value retval;
> > arraylist_object *intern;
> >
> > intern = emalloc(sizeof(arraylist_object));
> >
> > intern->size = 0;
> >
> > zend_object_std_init(&intern->std, ce TSRMLS_CC);
> >
> > ALLOC_INIT_ZVAL(intern->elements);
> > array_init(intern->elements);
> >
> > intern->std.ce = ce;
> >
> > object_properties_init(&intern->std, ce);
> >
> > retval.handle = zend_objects_store_put(intern, NULL,
> > destroy_arraylist_object, NULL TSRMLS_CC);
> > retval.handlers = &arraylist_object_handlers;
> >
> > return retval;
> > }
> > /* }}} */
> >
> > I'm not getting any compilation errors or any type of warnings.
> Everything
> > seems to be glueing itself correctly.
> >
> > I'm hoping someone can see what I'm doing wrong so that I can move
> forward
> > with this.
> >
> > thanks!
>
>
>
> --
> Laruence Xinchen Hui
> http://www.laruence.com/
>
