This also fixes skipped tests due to different naming "zend-test"
instead of "zend_test" and "PDO" instead of "pdo":
- ext/dom/tests/libxml_global_state_entity_loader_bypass.phpt
- ext/simplexml/tests/libxml_global_state_entity_loader_bypass.phpt
- ext/xmlreader/tests/libxml_global_state_entity_loader_bypass.phpt
- ext/zend_test/tests/observer_sqlite_create_function.phpt
EXTENSIONS section is used for the Windows build to load the non-static
extensions.
Closes GH-13276
This syncs the installed sapi and extension headers on *nix and Windows
systems by installing only what is intended outside of php-src.
- ext/gd: without gd_arginfo.h and gd_compat.h
- ext/hash: php_hash_joaat.h and php_hash_fnv.h added also on Windows
installation; xxhash/xxhash.h added on both installations as it is
included in php_hash_xxhash.h; Include path for xxhash.h changed to
relative so the php_hash_xxhash.h can be included outside of php-src;
Redundant include flags removed
- ext/iconv: without iconv_arginfo.h
- ext/mysqli: mysqli_mysqlnd.h was missing on Windows
- ext/phar: php_phar.h was missing on Windows
- ext/sodium: php_libsodium.h was missing on *nix
- ext/xml: without xml_arginfo.h
- sapi/cli: cli.h was missing on Windows
Closes GH-13210
Closes GH-13213
take2 of #12657
## About Firebird transaction
Firebird is a full transactional database, so the DB itself does not support autocommit mode. (Strictly, there is an autocommit mode, but it is a different concept from the "autocommit" that we are used to with MySQL and others.)
Therefore, a transaction must have started before any operation is performed, and autocommit mode had to be emulated in PHP.
I made sure that a transaction always exists when in autocommit mode. Since the `in_transacntion` function does not work as expected, I have introduced `H->in_manually_txn` to determine whether a transaction is being manually manipulated.
## There are two types of commit/rollback
(I'm not talking about two-phase commit. This change does not take into account two-phase commit.)
There are `isc_commit_retaining` which starts a transaction again in the same context immediately after committing, and `isc_commit_transaction` which closes the transaction as is.
Similarly, there are two types of rollback.
-----------
Due to the default value of the transaction isolation level, autocommit mode may obtain unintended results.
Regarding this, it would be too large to include support for transaction isolation levels in this PR, so I will leave it as is for now.
We may OOM during object initialization. In this case, free_obj needs to guard
against NULL values. There may be more cases where this is an issue, these were
the ones I was able to discover via script.
Fixes GH-11734
This also includes a fix for the MySQL ND driver to actually respect the user decided behaviour.
Closes GH-11622
Signed-off-by: George Peter Banyard <girgias@php.net>
This reverts commit 94ee4f9834.
The commit was a bit too late to be included in PHP 8.2 RC1. Given it's a massive ABI break, we decide to postpone the change to PHP 8.3.
Using php_info_print_table_header() for "Foo: bar" looks odd and out of place,
because the whole line is colored. It is also questionable from a HTML
semantics point of view, because it does not described the columns that follow.
The use of this across extensions is inconsistent. It was part of the skeleton,
but ext/date or ext/json already use a regular row.
This avoids a possible significant performance penalty, when some leaf function was observed, deep in the stack.
As a side effect, we are not iterating over prev_execute_data anymore and thus, non-observed fake frames, possibly on stack, cannot have any impact on the observer anymore (especially within zend_observer_fcall_end_all).
Saving the previous observer happens now directly on the VM stack. If there is any observer, function frames are allocated an extra zval (the last temporary), which will, on observed frames, contain the previous observed frame address.
There are two main motivations to this:
a) The logic for handling internal and userland observation can be unified.
b) Unwinding of observed functions on a bailout does notably not include observers. Even if users of observers were to ensure such handling themselves, it would be impossible to retain the relative ordering - either the user has to unwind all internal observed frames before the automatic unwinding (zend_observer_fcall_end_all) or afterwards, but not properly interleaved.
Signed-off-by: Bob Weinand <bobwei9@hotmail.com>
