Instead of handling shebang lines by adjusting the file pointer in
individual SAPIs, move the handling into the lexer, where this is
both a lot simpler and more robust. Whether the shebang should be
skipped is controlled by CG(skip_shebang) -- we might want to do
that in more cases.
This fixed bugs #60677 and #78066.
Keep track of delayed variance obligations and check them after
linking a class is otherwise finished. Obligations may either be
unresolved method compatibility (because the necessecary classes
aren't available yet) or open parent/interface dependencies. The
latter occur because we allow the use of not fully linked classes
as parents/interfaces now.
An important aspect of the implementation is we do not require
classes involved in variance checks to be fully linked in order for
the class to be fully linked. Because the involved types do have to
exist in the class table (as partially linked classes) and we do
check these for correct variance, we have the guarantee that either
those classes will successfully link lateron or generate an error,
but there is no way to actually use them until that point and as
such no possibility of violating the variance contract. This is
important because it ensures that a class declaration always either
errors or will produce an immediately usable class afterwards --
there are no cases where the finalization of the class declaration
has to be delayed until a later time, as earlier variants of this
patch did.
Because variance checks deal with classes in various stages of
linking, we need to use a special instanceof implementation that
supports this, and also introduce finer-grained flags that tell us
which parts have been linked already and which haven't.
Class autoloading for variance checks is delayed into a separate
stage after the class is otherwise linked and before delayed
variance obligations are processed. This separation is needed to
handle cases like A extends B extends C, where B is the autoload
root, but C is required to check variance. This could end up
loading C while the class structure of B is in an inconsistent
state.
Causes build failure on release+zts azure build. I'm rewriting this
code to separate the if/else handling, because they don't really
have anything in common anyway...
Even though we don't need it at runtime, add the BIND_IMPLICIT
flag to BIND_STATIC as well, so we can distinguish this case in
type inference.
This fixes a JIT miscompile in arrow_functions/002.phpt.
Fixes bug #76451, and more importantly lays necessary groundwork for
covariant/contravariant types. Bug #76451 is just an edge case, but
once covariance is introduced this will become a common problem instead.
Make sure to always fetch the RHS of a list assignment first, instead
of special casing known self-assignments, which will not detect cases
using references correctly.
As a side-effect, it is no longer possible to do something like
byRef(list($x) = $y). This worked by accident previously, but only
if $y was a CV and the self-assignment case did not trigger.
However it shouldn't work for the same reason that byRef($x = $y)
doesn't. Conversely byRef(list(&$x) = $y) and byRef($x =& $y)
continue to be legal.
This reverts commit e528762c1c.
Dmitry reports that this has a non-trivial impact on parsing
overhead, especially on 32-bit systems. As we don't have a strong
need for this change right now, I'm reverting it.
See also comments on
e528762c1c.
Locations for AST nodes are now tracked with the help of bison
location tracking. This is more accurate than what we currently do
and easier to extend with more information.
A zend_ast_loc structure is introduced, which is used for the location
stack. Currently it only holds the start lineno, but can be extended
to also hold end lineno and offset/column information in the future.
All AST constructors now accept a zend_ast_loc* as first argument, and
will use it to determine their lineno. Previously this used either the
CG(zend_lineno), or the smallest AST lineno of child nodes.
On the parser side, the location structure for a whole rule can be
obtained using the &@$ character salad.
This reverts commit a9e6667817.
Breakage found in the wild: Mockery uses a parent:: call in the
implementation regardless of whether the class has a parent or not:
4324afeaf9/library/Mockery/Mock.php (L600)
This change is not worth the compat break in 7.4.
The sprintf function has been normalized to php_sprintf via
61364b5bb1.
This patch removes the checks to make a custom sprintf function
The ZEND_BROKEN_SPRINTF has been removed and the
hardcoded #define zend_sprintf sprintf is used.
The php_sprintf and zend_sprintf are now symbols to sprintf.
This patch now removes the custom PHP definitions of the php_sprintf and
zend_sprintf functions in favor of the C99 sprintf which is also
standardized in C89 already. Once, on some systems sprintf didn't behave
in same way.
Opcode order changes in 7.4 and the EXT_STMT is now declare the
DECLARE_ANON. Fix this by returning the opline from compile_class_decl
to avoid any fragile opcode searching.
If someone has a better patch, please merge it ASAP, this appears to be correct as I and Nikita originally thought.
Revert "Revert "zend_get_call_op ignoring compiler flags zend_get_call_op will ignore ZEND_COMPILE_IGNORE_USER_FUNCTIONS and ZEND_COMPILE_IGNORE_USER_FUNCTIONS, breaking the intention of these flags""
This reverts commit 0bbbd0f9e7.
* PHP-7.4:
Revert "zend_get_call_op ignoring compiler flags zend_get_call_op will ignore ZEND_COMPILE_IGNORE_USER_FUNCTIONS and ZEND_COMPILE_IGNORE_USER_FUNCTIONS, breaking the intention of these flags"
* PHP-7.4:
zend_get_call_op ignoring compiler flags zend_get_call_op will ignore ZEND_COMPILE_IGNORE_USER_FUNCTIONS and ZEND_COMPILE_IGNORE_USER_FUNCTIONS, breaking the intention of these flags
This avoids writing this cache at runtime, which is illegal if
preloading is used.
Not every serialize/unserialize function actually belongs to the
Serializable interface, but I think it's not a problem to assign
these anyway -- whether they are used ultimately depends on whether
Serializable is implemented.
Alternatively it might make sense to just drop these entirely. I
don't think this is performance critical functionality.
This patch removes the so called local variables defined per
file basis for certain editors to properly show tab width, and
similar settings. These are mainly used by Vim and Emacs editors
yet with recent changes the once working definitions don't work
anymore in Vim without custom plugins or additional configuration.
Neither are these settings synced across the PHP code base.
A simpler and better approach is EditorConfig and fixing code
using some code style fixing tools in the future instead.
This patch also removes the so called modelines for Vim. Modelines
allow Vim editor specifically to set some editor configuration such as
syntax highlighting, indentation style and tab width to be set in the
first line or the last 5 lines per file basis. Since the php test
files have syntax highlighting already set in most editors properly and
EditorConfig takes care of the indentation settings, this patch removes
these as well for the Vim 6.0 and newer versions.
With the removal of local variables for certain editors such as
Emacs and Vim, the footer is also probably not needed anymore when
creating extensions using ext_skel.php script.
Additionally, Vim modelines for setting php syntax and some editor
settings has been removed from some *.phpt files. All these are
mostly not relevant for phpt files neither work properly in the
middle of the file.
Access to undefined constants will now always result in an Error
exception being thrown.
This required quite a few test changes, because there were many
buggy tests that unintentionally used bareword fallback in combination
with error suppression.
The only remaining case-insensitive constants are null, true and
false, which are handled explicitly.
In the future we may convert them from constants to reserved keywords.
There are probably some improvements we can do to the SPL
implementation now that __autoload() is gone. In particular having
EG(autoload_func) as a property zend function, rather than a simple
callback probably doesn't make sense.
This is useful for coverage. While it is currently safe to just
skip over the SWITCH_* opcodes, this may not be true in the future
due to opcache optimizations, so it's safer to disable emission of
SWITCH_* opcodes entirely.
RFC: https://wiki.php.net/rfc/null_coalesce_equal_operator
$a ??= $b is $a ?? ($a = $b), with the difference that $a is only
evaluated once, to the degree that this is possible. In particular
in $a[foo()] ?? $b function foo() is only ever called once.
However, the variable access themselves will be reevaluated.
Instead of interleaving creation of live-ranges with the main
compiler code, compute them in a separate pass over the opcodes
as part of pass_two. Additionally, do not keep live ranges
synchronized during optimization in opcache and instead use the
same mechanism to recompute them after optimization.
RFC: https://wiki.php.net/rfc/typed_properties_v2
This is a squash of PR #3734, which is a squash of PR #3313.
Co-authored-by: Bob Weinand <bobwei9@hotmail.com>
Co-authored-by: Joe Watkins <krakjoe@php.net>
Co-authored-by: Dmitry Stogov <dmitry@zend.com>
We already detect the case where we're entirely outside a class --
now also check whether there actually is a parent.
This is a minor BC break, in that code that was never executed
might have previously contained an invalid parent:: reference without
generating an error.
Instead of representing this as a ZEND_AST_CLASS_CONST with a
"class" constant name.
Class constants and ::class are unrelated features that happen to
share syntax, so represent and handle them separately.
Previously this triggered an assertion failure. The behavior is
not quite correct, in that self::class should generate an exception
if there is no self, but returns an empty string here. Fixing that
would be a bit too intrusive for the 7.2 branch.
This does not print the exact line of the comma, but rather the line
of the previous element. This should generally be "good enough", as
the line number is close (off by one) to the actual issue now.
Previously it would point to the start of the array, which may be
very far away.
