Remove EXT_TYPE_UNUSED in favor of IS_UNUSED

This means we no longer allocate an unused VAR for the retval of
instructions that support unused results.

Nearly all instructions a

Remove EXT_TYPE_UNUSED in favor of IS_UNUSED

This means we no longer allocate an unused VAR for the retval of
instructions that support unused results.

Nearly all instructions already used the result variable only if
it was used. The only exception to this was the return value
variable for internal function call results. I've adjusted the code
to use a stack zval for the unused return case now. As we have
retval specialization now, we know that it doesn't matter.

show more ...

Fixed bug #71441 (Typehinted Generator with return in try/finally crashes)

We don't need to dup zend_class_constant here.

quote why the internal class is still using duplication:
```
for internal classes, the zend_class_constant is malloc-ed. we need to

We don't need to dup zend_class_constant here.

quote why the internal class is still using duplication:
```
for internal classes, the zend_class_constant is malloc-ed. we need to
free it.

if (const->ce == ce) {
zval_ptr_dtor(&const->value);
free(const)
}
so, if two classes share one const, and it(parent class) was freed
before, this read(in child class, const->ce) is invalid..

and destroy_zend_class is called via zend_hash_destroy(class_table).
which is not in reverse order... so, parent classes are dtor first.

if we want this work, we should change that order.
```

show more ...

Trigger op_array_handler() if HAVE_OP_ARRAY_HANDLER is set

Duplicated i

bump year which is missed in rev 49493a2

bump year which is missed in rev 49493a2

Remove uses of VARs in extended_value

The DECLARE_(ANON_)INHERITED_CLASS(_DELAYED) opcodes were
referencing the parent ce VAR through extended_value. This is
hacky and we can't track

Remove uses of VARs in extended_value

The DECLARE_(ANON_)INHERITED_CLASS(_DELAYED) opcodes were
referencing the parent ce VAR through extended_value. This is
hacky and we can't track the def-use chain in SSA.

To avoid this, the layout of declaration opcodes is changed
as follows: op1 points to the lcname and rtd_key literals, in
that order. (For anon/lambda declarations only one of lcname or
rtd_key is present.) This frees up op2, which is now used to
reference the parent ce VAR in inheriting declarations. The
jmp offset for anon class declarations is moved frop op2 to
extended_value.

The changes were applied both to class and function declarations
to keep everything symmetric.

show more ...

Add result op for ASSERT_CHECK to fix SSA

Otherwise one source of the phi node at the join point will be
borked.

I'm marking the ASSERT_CHECK result as unused in pass2, which is

Add result op for ASSERT_CHECK to fix SSA

Otherwise one source of the phi node at the join point will be
borked.

I'm marking the ASSERT_CHECK result as unused in pass2, which is
a bit ugly.

show more ...

Implemented the RFC `Support Class Constant Visibility`.

Squashed commit of the following:

commit f11ca0e7a57793fa0e3e7f6d451720e6c42bb0b9
Author: Dmitry Stogov <dmitry@zend.com

Implemented the RFC `Support Class Constant Visibility`.

Squashed commit of the following:

commit f11ca0e7a57793fa0e3e7f6d451720e6c42bb0b9
Author: Dmitry Stogov <dmitry@zend.com>
Date: Tue Dec 8 12:38:42 2015 +0300

Fixed test expectation

commit 211f873f542504d0a0f72b6b5cb23908a1c99a2d
Author: Dmitry Stogov <dmitry@zend.com>
Date: Tue Dec 8 12:28:38 2015 +0300

Embed zend_class_constant.flags into zend_class_constants.value.u2.access_flags

commit 51deab84b2cdbf9cdb1a838cf33b2ee45c61748b
Author: Dmitry Stogov <dmitry@zend.com>
Date: Mon Dec 7 11:18:55 2015 +0300

Fixed issues found by Nikita

commit 544dbd5b47e40d38a8ccb96bc5583e9cb7fdd723
Author: Dmitry Stogov <dmitry@zend.com>
Date: Sat Dec 5 02:41:05 2015 +0300

Refactored immplementation of https://wiki.php.net/rfc/class_const_visibility
@reeze created an RFC here and I emailed internals here and didn't get any responses positive/negative.

show more ...

Squashed commit of the following:

commit afe963e6cc289696e60c6c679796ba2197c52b3b
Author: Dmitry Stogov <dmitry@zend.com>
Date: Fri Nov 13 15:32:29 2015 +0300

Added ne

Squashed commit of the following:

commit afe963e6cc289696e60c6c679796ba2197c52b3b
Author: Dmitry Stogov <dmitry@zend.com>
Date: Fri Nov 13 15:32:29 2015 +0300

Added news entry

commit a126b891c97848dd7ef8f1abf716328c46e0f19c
Author: Dmitry Stogov <dmitry@zend.com>
Date: Fri Nov 13 15:29:21 2015 +0300

VERIFY_RETURN_TYPE doesn't have to cleanup operand on exception, bacause now, live temporary variables are released by exception unwinder.

commit 0db475e98786e6bcaa8401ee3e0b33743b9a2f2b
Author: Dmitry Stogov <dmitry@zend.com>
Date: Thu Nov 12 22:55:39 2015 +0300

Fixed copy/paste

commit 0ac73fe7174bec9de9a610319a98b259bea67f7f
Author: Dmitry Stogov <dmitry@zend.com>
Date: Wed Nov 11 16:11:50 2015 +0300

Fixed bug #62210 (Exceptions can leak temporary variables)

show more ...

We don't nees zend_op_array->brk_cont_array at run-time anymore.
Move zend_op_array->brk_cont_array into CG(context).brk_cont_array.
Use more compact zend_op_array->live_range instead of zend

We don't nees zend_op_array->brk_cont_array at run-time anymore.
Move zend_op_array->brk_cont_array into CG(context).brk_cont_array.
Use more compact zend_op_array->live_range instead of zend_op_array->brk_cont_array.
Semantic is kept unchanged.

show more ...

Changed meaning of "op2" for ZEND_FREE, ZEND_FE_FREE, ZEND_FAST_CALL, ZEND_FAST_RET.
Previously it was an instruction number.
Now it's an index in op_array->try_cacth_array[].

Changed ZEND_CLONE->extended_value meaning to relative offset (previously it was absolute opline number)

Extended zend_extension API to allow storing additional data associated with op_arrays in opcache SHM. (op_array->reserved[])

Add myself into list of authors of the most refactored files.

Removed unused functions

Get rid of ZEND_FAST_CALL_UNBOUND

Move most "finally" related code-generation from pass_two() to compiler.

Try to fix finally issue

Fix valgrind errors in phpdbg

Revert "We cannot safely assume that all op array will be refcount 0 after execution"
This reverts commit b6936adb58288a0606ed847802d9226cddb41e2b.

Fix valgrind errors in phpdbg

Revert "We cannot safely assume that all op array will be refcount 0 after execution"
This reverts commit b6936adb58288a0606ed847802d9226cddb41e2b.

This change turns out to not have been a clever idea and was causing more weirdness than it helped...

show more ...

Fully fix typed generator returns

Fix bug using not set temporary with typed generator returns

Cleanup shutdown, enable proper memory leak displaying
phpdbg should not memory leak...

Opcache is assuming that op_arrays without refcount are always efree()d
We maybe should search for some other way, because just not refcounting the op_array is rather a dirty hack.
If possibl

Opcache is assuming that op_arrays without refcount are always efree()d
We maybe should search for some other way, because just not refcounting the op_array is rather a dirty hack.
If possible, we should change opcache and just expect op_array->refcount != NULL ... but I have not found a clean way yet.

show more ...