Squashed commit of the following:
commit 0361dbe356
Author: Andrea Faulds <ajf@ajf.me>
Date: Fri Mar 25 16:59:20 2016 +0000
UPGRADING and NEWS
commit dca9d4a36c
Author: Andrea Faulds <ajf@ajf.me>
Date: Fri Mar 25 16:45:18 2016 +0000
Add tests contributed by @jesseschalken
commit e557f77eab
Author: Andrea Faulds <ajf@ajf.me>
Date: Fri Mar 25 16:44:51 2016 +0000
Rebuild VM
commit 70942e4c3c
Author: Andrea Faulds <ajf@ajf.me>
Date: Wed Feb 24 13:12:26 2016 +0000
Add test for evaluation order of nested list() keys
commit ed3592e80c
Author: Andrea Faulds <ajf@ajf.me>
Date: Wed Feb 24 12:42:04 2016 +0000
Add test for evaluation order
commit 589756cbcc
Author: Andrea Faulds <ajf@ajf.me>
Date: Tue Jan 19 17:29:34 2016 +0000
Allow arbitrary expressions for key
commit 3f622077c3
Author: Andrea Faulds <ajf@ajf.me>
Date: Tue Jan 19 17:45:10 2016 +0000
Remove compile-time HANDLE_NUMERIC (see bug #63217)
commit bab758119a
Author: Andrea Faulds <ajf@ajf.me>
Date: Sun Jan 17 01:20:26 2016 +0000
Handle numeric strings
commit 14bfe93ddc
Author: Andrea Faulds <ajf@ajf.me>
Date: Sun Jan 17 01:09:36 2016 +0000
Allow trailing comma
commit f4c8b2cb30
Author: Andrea Faulds <ajf@ajf.me>
Date: Sat Jan 16 23:47:11 2016 +0000
Add tests
commit 0085884a61
Author: Andrea Faulds <ajf@ajf.me>
Date: Sat Jan 16 22:24:23 2016 +0000
Handle non-integer/string opcodes
commit e572d2d0ad
Author: Andrea Faulds <ajf@ajf.me>
Date: Sat Jan 16 21:10:33 2016 +0000
Disallow mixing keyed and unkeyed list() elements
commit cede13ccfe
Author: Andrea Faulds <ajf@ajf.me>
Date: Sun Jan 10 20:46:44 2016 +0000
list() with keys (no foreach or tests)
While the def starts at DO_FCALL, the variable should still be the
result of NEW, not DO_FCALL.
I had to fix the test for #68652, because the code started to
(correctly) free the "new self()" object, which triggered an
infinite destructor loop.
The current yield from implementation doesn't support by-ref
yields. It's likely not worthwhile to actually implement this,
but we should at least keep the door open for the future by
issuing a compile error.
Refs bug #71252.
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.
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.
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.
This does not collide with NO_AUTOLOAD -- missing classes will be
silenced, but invalid use of self etc will result in an exception
instead of a fatal error.
Just found this old AST todo...
It no longer makes sense as an optimization due to the introduction
of BIND_GLOBAL, however it ensures that the global variable name is
not evaluated twice, which is presumably something we wish to
guarantee.
