Internally accessible via zend_jit_blacklist_function / externally via opcache_jit_blacklist.
The functionality currently only affects tracing JIT, but may be extended to other JIT modes in future.
Note that this is not actually security related[1], but still a
reasonable sanity check.
"If a function be advertised to return an error code in the event of
difficulties, thou shalt check for that code, yea, even though the
checks triple the size of thy code and produce aches in thy typing
fingers, for if thou thinkest it cannot happen to me, the gods shall
surely punish thee for thy arrogance." – Henry Spencer
[1] <https://github.com/libgd/libgd/issues/697#issuecomment-2369613187>
libcurl ref: [`CURLINFO_POSTTRANSFER_TIME_T`](https://curl.se/libcurl/c/CURLINFO_POSTTRANSFER_TIME_T.html)
`CURLINFO_POSTTRANSFER_TIME_T` is a libcurl info option that returns
the time it took to "post" the transfer. Available since libcurl 8.10.0
This value is also exposed as `posttransfer_time_us` in the
`curl_getinfo()` return value when the `$option` parameter is not
passed.
PQclosePrepared allows the statement's name to be reused thus allowing
cache solutions to work properly ; whereas, for now, the `DEALLOCATE
<statement>` query is used which free entirely the statement's resources.
close GH-13316
We need to avoid signed integer overflows which are undefined behavior.
We catch that, and set `offset` to `ZEND_LONG_MAX` (which is also the
largest value of `zend_off_t` on all platforms). Of course, that seek
may fail, but even if it succeeds, the stream is no longer readable,
but that matches the current behavior for offsets near `ZEND_LONG_MAX`.
Closes GH-15989.
This adds support for `CURLOPT_DEBUGFUNCTION`[^1] Curl option to set a
custom callback that gets called with debug information during the
lifetime of a Curl request.
The callback gets called with the `CurlHandle` object, an integer
containing the type of the debug message, and a string containing the
debug message. The callback may get called multiple times with the
same message type during a request.
PHP already uses `CURLOPT_DEBUGFUNCTION` functionality to internally
to expose a Curl option named `CURLINFO_HEADER_OUT`.
However,`CURLINFO_HEADER_OUT` is not a "real" Curl option supported
by libcurl. Back in 2006, `CURLINFO_HEADER_OUT` was added[^2] as
a Curl option by using the debug-callback feature. Git history does
not run that back to show why `CURLINFO_HEADER_OUT` was added as a
Curl option, and why the other debug types (such as
`CURLINFO_HEADER_IN` were not added as Curl options, but this seems
to be a historical artifact when we added features without trying
to be close to libcurl options.
This approach has a few issues:
1. `CURLINFO_HEADER_OUT` is not an actual Curl option supported by
upstream libcurl.
2. All of the Curl options have `CURLOPT_` prefix, and `CURLINFO_HEADER_OUT`
is the only Curl "option" that uses the `CURLINFO` prefix. This exception
is, however, noted[^3] in docs.
3. When `CURLINFO_HEADER_OUT` is set, the `CURLOPT_VERBOSE` is also implicitly
set. This was reported[^4] to bugs.php.net, but the bug is marked as wontfix.
This commit adds support for `CURLOPT_DEBUGFUNCTION`. It extends the existing
`curl_debug` callback to store the header-in information if it encounters
a debug message with `CURLINFO_HEADER_OUT`. In all cases, if a callable
is set, it gets called.
`CURLOPT_DEBUGFUNCTION` intends to replace `CURLINFO_HEADER_OUT` Curl
option as a versatile alternative that can also be used to extract
other debug information such as SSL data, text information messages,
incoming headers, as well as headers sent out (which `CURLINFO_HEADER_OUT`
makes available).
The callables are allowed to throw exceptions, but the return values are
ignored.
`CURLOPT_DEBUGFUNCTION` requires `CURLOPT_VERBOSE` enabled, and setting
`CURLOPT_DEBUGFUNCTION` does _not_ implicitly enable `CURLOPT_VERBOSE`.
If the `CURLOPT_DEBUGFUNCTION` option is set, setting `CURLINFO_HEADER_OUT`
throws a `ValueError` exception. Setting `CURLOPT_DEBUGFUNCTION` _after_
enabling `CURLINFO_HEADER_OUT` is allowed. Technically, it is possible
for both functionality (calling user-provided callback _and_ storing
header-out data) is possible, setting `CURLINFO_HEADER_OUT` is not
allowed to encourage the use of `CURLOPT_DEBUGFUNCTION` function.
This commit also adds the rest of the `CURLINFO_` constants used as
the `type` integer value in `CURLOPT_DEBUGFUNCTION` callback.
---
[^1]: [cur.se - CURLOPT_DEBUGFUNCTION](https://curl.se/libcurl/c/CURLOPT_DEBUGFUNCTION.html)
[^2]: [`5f25d80`](5f25d80d10)
[^3]: [curl_setopt doc mentioning `CURLINFO_` prefix is intentional](https://www.php.net/manual/en/function.curl-setopt.php#:~:text=prefix%20is%20intentional)
[^4]: [bugs.php.net - `CURLOPT_VERBOSE` does not work with `CURLINFO_HEADER_OUT`](https://bugs.php.net/bug.php?id=65348)
Exactly these tests are failing on all our macOS x64 CI runners for
about a week. For now we mark them as xfail, to get back to a green
CI pipeline.
Closes GH-16002.
This effectively inlines the behaviour of php_mkdir_ex() which is a deprecated API from at least 17 years ago, and also fixes some of the return values.
This also removes a dependency on ext/standard
We need to avoid signed integer overflows which are undefined behavior.
We catch that, and set `offset` to `ZEND_LONG_MAX` (which is also the
largest value of `zend_off_t` on all platforms). Of course, after such
a seek a stream is no longer readable, but that matches the current
behavior for offsets near `ZEND_LONG_MAX`.
Closes GH-15989.
This works, but UBSan running on a big endian platform (in this, ppc64)
will complain that the ((uchar*)buffer)[n] is int, and shifting that
could be weird. Since the value of PHAR_GET_32 et al are almost always
unsigned, it makes sense to cast these as unsigned.
Fixes phar tests on a big endian system with UBSan enabled.
Instead of always saying that a name is reserved or deprecated and
cannot/should not be used as a class name, take the usage into account and say
the name cannot be used as an enum name, trait name, etc. In the process, for
class names add a missing "a".
The actual problem is our `shmget()` implementation which does not care
to set `errno` appropriately; that should be fixed, although mapping
the error conditions to those specified by POSIX might be hard.
For now, we only make the test case more resilient by ignoring the
exact error; "No error" doesn't make sense anyway.
This was first reported as a leak in GH-15026, but was mistakingly
believed to be a false positive. Then an assertion was added and it got
triggered in GH-15908. This fixes the leak. Upon merging into master the
assertion should be removed as well.
Closes GH-15924.
This test queries `extOutput` for the first line of output of the
command which was executed when snmpd started (`HexTest` in snmpd.conf).
Since there is no `/bin/sh` on Windows, no command would be run, and as
such we received an empty string. We fix that by dynamically adjusting
snmpd.conf to run a JScript which has the same output as bigtest.
We also make the test diff more helpful in case of failures, where so
far we only would have known that the output had a different message
digest.
As of net-snmp 5.8.0, the library defines their own `(v)asprintf()` if
not available on the system. However, PHP also does this, so when
building ext/snmp there would be conflicting declarations on Windows.
To avoid this, we explictly define `HAVE_ASPRINTF`, so net-snmp does
not redeclare when its headers are included.
SplFixedArray should've never get supported in ArrayObject because it's
overloaded, and so that breaks assumptions. This regressed in c4ecd82f.
Closes GH-15947.
We can use `memcmp()` directly and skip some of the logic handling
in `zend_binary_strcmp()`. `perf record` shows a reduction for
`dom_html5_serializes_as_void()` from 3.12% to 0.77%.
