Closes GH-16730

* PHP-8.4:
  UPGRADING: Fix `CURLOPT_TCP_KEEPCNT` entry

* PHP-8.4:
  Fix i386 release build warning

When returning an UNDEF value, it actually becomes NULL.
The following code took this into account:
https://github.com/php/php-src/blob/28344e0445bc2abae8dc5f1376aa0ff350e6d66d/ext/opcache/jit/zend_jit_trace.c#L2196-L2199

But the stack does not update the type to NULL, causing a mismatch.

Closes GH-16784.

Co-authored-by: Dmitry Stogov <[email protected]>

* PHP-8.2:
  Fix GH-16770: Tracing JIT type mismatch when returning UNDEF

* PHP-8.3:
  Fix GH-16770: Tracing JIT type mismatch when returning UNDEF

* PHP-8.4:
  Fix GH-16770: Tracing JIT type mismatch when returning UNDEF

PHP-8.1 is not ready for ICU >= 75.1 which requires C++17 support.
Thus we force the usage of icu4c@74, what is scheduled for removal in
May 2025, though.

Closes GH-16789.

* PHP-8.1:
  Stick with icu4c 74.2 on macOS CI for PHP-8.1

* PHP-8.2:
  Stick with icu4c 74.2 on macOS CI for PHP-8.1

* PHP-8.3:
  Stick with icu4c 74.2 on macOS CI for PHP-8.1

* PHP-8.4:
  Stick with icu4c 74.2 on macOS CI for PHP-8.1

…all.

close GH-16813

FreeBSD 13.2 is no longer supported[1], and apparently the respective
Cirrus CI image has been removed.  We update to FreeBSD 13.3.

This also requires to disable some compiler warnings now, and adapt a
test case.

[1] <https://www.freebsd.org/security/#sup>
[2] <https://github.com/php/php-src/runs/33044888209>

Co-authored-by: David Carlier <[email protected]>

Closes GH-16817.

* PHP-8.1:
  Update FreeBSD CI

* PHP-8.2:
  Update FreeBSD CI

* PHP-8.3:
  Update FreeBSD CI

* PHP-8.4:
  Update FreeBSD CI

Closes GH-16788.

* PHP-8.2:
  Adapt ext/intl tests for ICU 75.1 and 76.1

* PHP-8.3:
  Adapt ext/intl tests for ICU 75.1 and 76.1

* PHP-8.4:
  Adapt ext/intl tests for ICU 75.1 and 76.1

zend_is_callable_ex() can unfortunately emit a deprecation, and then
a user error handler can throw an exception. This causes an assert
failure at ZEND_VM_NEXT_OPCODE(). We fix this by checking if there's an
exception after zend_is_callable_ex().

Closes GH-16803.

* PHP-8.2:
  Fix GH-16799: Assertion failure at Zend/zend_vm_execute.h:7469

* PHP-8.3:
  Fix GH-16799: Assertion failure at Zend/zend_vm_execute.h:7469

* PHP-8.4:
  Fix GH-16799: Assertion failure at Zend/zend_vm_execute.h:7469

* PHP-8.4:
  Fix test expectation for different closure name in 8.4

And fix a memleak while here.

Closes GH-16804.

* PHP-8.3:
  Fix GH-16802: open_basedir bypass using curl extension

* PHP-8.4:
  Fix GH-16802: open_basedir bypass using curl extension

There is no such dependency; only libxml2 depends on libiconv.  So when
php_libxml.dll is built, it needs to be linked against libiconv, or,
when ext/iconv has been configured as static extension, against
php8.dll.

…t() with a xml element input

When the current data is invalid, NULL must be returned. At least that's
how the check in SPL works and how other extensions do this as well.
If we don't do this, an UNDEF value gets propagated to a return value
(misprinted as null); leading to issues.

Closes GH-16825.

* PHP-8.2:
  Fix GH-16808: Segmentation fault in RecursiveIteratorIterator->current() with a xml element input

…is in a document causes UAF

Closes GH-16824.

* PHP-8.3:
  Fix GH-16777: Calling the constructor again on a DOM object after it is in a document causes UAF
  Fix GH-16808: Segmentation fault in RecursiveIteratorIterator->current() with a xml element input

* PHP-8.4:
  Fix GH-16777: Calling the constructor again on a DOM object after it is in a document causes UAF
  Fix GH-16808: Segmentation fault in RecursiveIteratorIterator->current() with a xml element input

Closes GH-16755

* PHP-8.4:
  ext/pdo_pgsql: Remove new PDO class constant specific to PGSQL driver

close GH-16836

I validated there were no spec changes for this, so we can bump the last
check date.

… OOB writes

…ing OOB writes

This fixes issues causing buffer over-read that leak heap content:
- RESP packet field default left over for COM_LIST
- RESP packet upsert filename
- OK packet message
- RESP packet for stmt row data
  - ps_fetch_from_1_to_8_bytes
  - ps_fetch_float
  - ps_fetch_double
  - ps_fetch_time
  - ps_fetch_date
  - ps_fetch_datetime
  - ps_fetch_string
  - ps_fetch_bit
- RESP packet for query row data (just possible overflow on 32bit)

It also adds various protocol tests using a new fake server.

Move the bound check upwards. Since this doesn't generate output we can
check the bound first.

…aarch64

* PHP-8.2:
  Fix GH-16829: Segmentation fault with opcache.jit=tracing enabled on aarch64

* PHP-8.3:
  Fix GH-16829: Segmentation fault with opcache.jit=tracing enabled on aarch64

* PHP-8.4:
  Fix GH-16829: Segmentation fault with opcache.jit=tracing enabled on aarch64

* PHP-8.2:
  Fixed test

* PHP-8.3:
  Fixed test

* PHP-8.4:
  Fixed test

The zend_hash_update_ind() variant unwraps indirects, rather than creating them.
Don't use _zend_hash_append_ind() because the property might already exist.

Fixes GH-16725
Closes GH-16805

* PHP-8.4:
  Fix get_object_vars() for non-hooked props in hooked prop iter

Closes GH-16844

* PHP-8.4:
  JIT: Set valid EX(opline) before calling gc_possible_root()

Not all of these are straight-forward to understand.

zend_save_lexical_state() can be nested multiple times, for example for
the parser initialization and then in the heredoc lexing. The input
should not be freed if we restore to the same filtered string.

Closes GH-16716.

* PHP-8.2:
  Fix GH-16630: UAF in lexer with encoding translation and heredocs

* PHP-8.3:
  Fix GH-16630: UAF in lexer with encoding translation and heredocs

* PHP-8.4:
  Fix GH-16630: UAF in lexer with encoding translation and heredocs

The official Windows builds and CI are doing this for ages, so it
appears to be overdue to finally switch the actual default.

Closes GH-16300.

This information can be occasionally useful, and would otherwise need
to be parsed from `phpinfo()` output.

However, maybe more importantly we unify the build date between what is
given by `php -v` and `php -i`, since these compilation units are not
necessarily preprocessed within the same second.

Closes GH-16747.

OPcache JIT does not support Windows ARM64, so we should not allow
`--enable-opcache-jit` in the first place.

Due to the way `ARG_ENABLE()` is handled on Windows, we do not attempt
to suppress the configure option, but just do not enable JIT when the
user attempts to, and adapt the help text.

Closes GH-16841.

* PHP-8.4:
  Fix GH-16839: Error on building Opcache JIT for Windows ARM64

See <#16747 (review)>.

Co-authored-by: Ayesh Karunaratne <[email protected]>

…t() (#16858)

* PHP-8.2:
  Backport JIT fix: set valid EX(opline) before calling gc_possible_root() (#16858)

* PHP-8.3:
  Backport JIT fix: set valid EX(opline) before calling gc_possible_root() (#16858)

* PHP-8.4:
  Backport JIT fix: set valid EX(opline) before calling gc_possible_root() (#16858)

… powers of 0. (#16694)

Closes #16694
Fixes #16236

* PHP-8.4:
  Fixed a bug in BcMath\Number::pow() and bcpow() when raising negative powers of 0. (#16694)

* PHP-8.4:
  [skip ci] NEWS for #16694