This header declaration is never supposed to be interpreted as
definition; otherwise, the handlers are not properly initialized, what
happens, for instance, with ASan instrumented MSVC builds.
Closes GH-7827.
Pass1 handles a number of special functions that can be evaluated
under some circumstances. Move the core logic into a separate
helper, as I believe that SCCP should reuse this.
Like `hash_file()`, `hash_hmac_file()` expects a filename, and not some
string data. Fixing this now, constitutes a (hopefully small) BC break
though.
Closes GH-7828.
These can be handled by the generic code. Worth noting that count
will usually go through ZEND_COUNT, and chr on constants is
evaluated in the compiler, so these are not particularly compile-time
sensitive either.
This handles references to the current class through its name
rather than self (and for cases where is is not linked yet and
thus not covered by the context lookup). Rather than handling this
only for FETCH_CLASS_CONSTANT optimization, integrate this into
the generic get_class_entry() utility.
Nowadays self::X is represented using an UNUSED operand with
FETCH_CLASS_SELF flag rather than a separate FETCH_CLASS instruction.
The code already handles the new pattern.
These are supported as constants nowadays, so we can drop the
string check.
Also fix a potential leak, though I believe this doesn't matter in
current usage, as it will effectively be suppressed during persist.
The generic code was rejecting this to go into a special code path
in SCCP. We should directly do that in SCCP instead, to still allow
the generic (and valid) replacement.
While we can't replace the instanceof operand, we will evaluate
the instanceof to false and replace its result anyway. Even in
cases where the instanceof user cannot be replaced, we already
have generic code to convert the opcode to QM_ASSIGN in that
case.
While the `path` is not supposed to be `NULL` for normal operation, it
is possible to create `SplFileInfo` objects where that is the case, and
we must not follow the null pointer.
Closes GH-7814.
On NetBSD, ifconf.ifc_buf member, unlike most of platforms, is a void
pointer.
We also fix the cpuinfo declarations with empty parameter lists.
Closes GH-7819.
This is a tiny improvement to run-tests.php which helps keep it from
tripping over hidden files, including autosave files from various
editors, metadata on certain platforms, etc.
Closes GH-7783.
Alex Dowad