year++

Fix DFG construction for VERIFY_RETURN

`use` only contains uses prior to definition. This was not
honoured for VERIFY_RETURN with a temporary operand.

The test case only breaks

Fix DFG construction for VERIFY_RETURN

`use` only contains uses prior to definition. This was not
honoured for VERIFY_RETURN with a temporary operand.

The test case only breaks on PHP 7.2.

show more ...

Update copyright headers to 2017

Fixed reference-counter inference

Updated reference-counting inferece rules.
The previous rules refected PHP-5 behavior and were completely wrong for PHP-7. (e.g. scalars don't have reference counters at all).
This change sho

Updated reference-counting inferece rules.
The previous rules refected PHP-5 behavior and were completely wrong for PHP-7. (e.g. scalars don't have reference counters at all).
This change shouldn't affect anything, because results of this inference are not used yet.

show more ...

Fixed compilation warnings

Split ZEND_SEND_VAR_NO_REF into ZEND_SEND_VAR_NO_REF and ZEND_SEND_VAR_NO_REF_EX (similar to ZEND_SEND_VAL) and remove ZEND_ARG_* flags.

This is not a problem here

if only for PHI placement, we may get rid of this `use` here.
But as the comment said, it is useful if we are going to do code generation

Replace BB end with BB len

And support empty blocks everywhere.

Merge def and gen sets

For live-variable analysis it does not matter if def includes
variables that are previously use in the same block, the data flow
equations still have the same

Merge def and gen sets

For live-variable analysis it does not matter if def includes
variables that are previously use in the same block, the data flow
equations still have the same result. As such there is no need to
compute separate gen & def sets.

I'm keeping the name "def", because use of "gen" in this context is
pretty confusing (gen is usually the use set, not the def set).

show more ...

Use return type hints for type inference and eliminate useless VERIFY_RETRUN_TYPE opcodes.

Manual CSE

Fix build

Combine conditions

Removed wrong (old) code

Remove dead code

OP_DATA isn't used in that way anymore

Respect RC_INFERENCE during DFG construction

To avoid inserting phis that are only relevant with rc inference
enabled. Suprisingly, this affects only 0.8% of phis.

Fix SSA for ZEND_YIELD

Yield-by-ref defs a ref var, yield-by-var only defs an rc1/rcn var
if rc inference is used.

Also move BIND_LEXICAL where it belongs in DFG construction.

Introduced BIND_STATIC opcode instead of FETCH_R/FETCH_W(static)+ASSIGN/ASSIGN_REF (similar to BIND_GLOBAL).
In the future we may refer to static variable by index instead of name, to eliminate h

Introduced BIND_STATIC opcode instead of FETCH_R/FETCH_W(static)+ASSIGN/ASSIGN_REF (similar to BIND_GLOBAL).
In the future we may refer to static variable by index instead of name, to eliminate hash lookup.

show more ...

Use worklist for DFG construction

About 40x faster.

Happy new year (Update copyright to 2016)

Introduce BIND_LEXICAL

This opcodes inserts a local CV into the closure static variable
table. This replaces the previous mechanism of having static
variables marked as LEXICAL, whic

Introduce BIND_LEXICAL

This opcodes inserts a local CV into the closure static variable
table. This replaces the previous mechanism of having static
variables marked as LEXICAL, which perform a symtable lookup
during copying.

This means a) functions which contain closures no longer have to
rebuild their symtable (better performance) and b) we can now track
used variables in SSA.

show more ...

Don't reuse SSA var in UNSET_VAR

Instead use the SSA var that UNSET_VAR actually defines. Otherwise
we get issues trying to DCE unsets.

Added e-SSA based DFA optimisation framework (incomplete)

Added e-SSA based DFA optimisation framework (incomplete)