The assertion failure was triggered in a debug code-path that validates
property types for internal classes.
zend_verify_internal_read_property_type was called with retval being a
reference, which is not allowed because that function eventually calls to
i_zend_check_property_type, which does not expect a reference.
The non-debug code-path already takes into account that retval can be a
reference, as it optionally dereferences retval.
Add a dereference in zend_verify_internal_read_property_type just before
the call to zend_verify_property_type, which is how other callers often
behave as well.
I learned this trick for doing a faster bounds check with both upper
and lower bounds by reading a disassembler listing of optimized code
produced by GCC; instead of doing 2 compares to check the upper and the
lower bound, add an immediate value to shift the range you are testing
for to the far low or high end of the range of possible values for the
type in question, and then a single compare will do. Intstead of
compare + compare + AND, you just do ADD + compare.
From microbenchmarking on my development PC, this makes strtoupper()
about 10% faster on long strings (~10,000 bytes).
Previously function JIT for the whole script (opcache.jit=1205) dumped
all op_arrays before JIT-ing the first functon. This complicated debugging
of particular JIT-ed function and leaded to usage of opcache.jit=1204.
Now both options should produce similar op_array/disasm interlived
output.
When this INI option is enabled, it reverts the line separator for
headers and message to LF which was a non conformant behavior in PHP 7.
It is done because some non conformant MTAs fail to parse CRLF line
separator for headers and body.
This is used for mail and mb_send_mail functions.
* Revert "ext/opcache: use C11 atomics for "restart_in" (#10276)"
This reverts commit 061fcdb0a5.
While the commit does indeed improve performance, @dstogov complained
that it disables the code path calling kill_all_lockers(), and thus
hanging workers are never killed and restarted.
https://github.com/php/php-src/pull/10276#issuecomment-1383593046
The fact that this feature was not implemented in the existing atomic
code path (i.e. Windows) did not mean that the feature was considered
not strictly necessary, but that the Windows implementation just did
not need the feature because SAPIs that work on Windows do not manage
child processes
https://github.com/php/php-src/pull/10276#issuecomment-1383868696https://github.com/php/php-src/pull/10276#issuecomment-1384235011
* ext/opcache: document lack of kill_all_lockers() on Windows
kill_all_lockers() is not implemented on Windows, and does not need to
be.
Upon freeing libxslt's context, every document which is not the *main*
document will be freed by libxslt. If a node of a document which is not
the main document gets returned to userland, we'd free the node twice:
- first by the cleanup of the xslt context
- and then by our own refcounting mechanism.
This was reported in bug 49634, and was fixed by always copying the
node (and later re-fixed in bug 70078).
The original fix is not entirely correct unfortunately because of the
following two main reasons:
- modifications to the node will only modify the copy, and not the original
- accesses to the parent, path, ... will not work
This patch fixes it properly by only copying the node if it origins from
a document other than the main document.
Co-authored-by: juha.ikavalko@agentit.fi
Closes GH-10318.
Thanks to the GitHub user 'titanz35' for pointing out that the new
implementation of mb_detect_encoding had poor detection accuracy on
UTF-8 and UTF-16 strings with a byte-order mark.
This file is unused (at least as of PHP 8.0.0); we also remove the
script that could generate it, and the respective info from
CONTRIBUTING.md.
Closes GH-10363.
The initialization routine for XXH3 was not prepared for exceptions from seed.
Fix this by using try_convert_to_string.
For discussion, please see: GH-10305
Closes GH-10352
Signed-off-by: George Peter Banyard <girgias@php.net>
The new SSE2-based implementation of mb_check_encoding for UTF-8 is
about 10% faster for 0-5 byte strings, more than 3 times faster for
~100-byte strings, and just under 4 times faster for ~10,000-byte
strings.
I believe it may be possible to make this function much faster again.
Some possible directions for further performance optimization include:
• If other ISA extensions like AVX or AVX-512 are available, use a
similar algorithm, but process text in blocks of 32 or 64 bytes
(instead of 16 bytes).
• If other SIMD ISA extensions are available, use the greater variety
of available instructions to make some of the checks tighter.
• Even if only SSE/SSE2 are available, find clever ways to squeeze
instructions out of the hot path. This would probably require a lot
of perusing instruction mauals and thinking hard about which SIMD
instructions could be used to perform the same checks with fewer
instructions.
• Find a better algorithm, possibly one where more checks could be
combined (just as the current algorithm combines the checks for
certain overlong code units and reserved codepoints).
This code path was only triggered if inst->cd == NULL. But the freeing
only happens if inst->cd != NULL. There is nothing to free here, so
remove this code. In fact, let's get rid of the goto too to make the
code more clear to read.
