Andrei Zmievski [Tue, 11 Jul 2000 16:48:03 +0000 (16:48 +0000)]
@- Improved array_multisort() to be able to specify sort type as well sort
@ order. Incidentally, it can be used to sort multi-dimensional arrays
@ as well. (Andrei)
Disable the hash_apply() protection on hashes that persist across requests - it's unsafe
because we may be aborted at any point
@- Fixed a possible data corruption in case of a huge amount of aborted requests (Zeev)
Danny Heijl [Mon, 10 Jul 2000 19:08:27 +0000 (19:08 +0000)]
- Fix the case of NULL BYTE and TEXT columns. (Danny).
ESQL/C (sometimes ?) does not set the indicator variable,
but the locator->loc_indicator instead.
Lift the add-once limit for libraries, because there are valid reasons
against it. When dealing with static libraries, the link order is
important and it might be necessary to add a library multiple times to
satisfy mutual symbol requirements.
In your extension's config.m4, you need to use the m4 macro
PHP_REQUIRE_CXX
This shall be called only, if your extension is enabled.
C++ source files shall not be put into the LTLIBRARY_SOURCES macro.
Instead, the filename of the object shall be put into
LTLIBRARY_OBJECTS_X. I.e. if your extension foo consists of foo.cxx, use
(ibase_blob_import) It's good to have link handle available, if you want to
use it...
(_php_ibase_exec) On error situations, only xsqlda was freed, not its
contents.
- I must be getting demented. Forgot to add and register list destructor
function when creating new list entry type. Several other modifications
too related to this. Blobs still not working.
* php4/ext/sockets.c, php4/ext/php_sockets.h:
Numerous changes. Many prototypes changed to be more like the
appropriate *NIX counterparts. Many new prototypes defining many
more advanced socket routines. Better AF_UNIX socket support.
bind() now recognizes the socket type and acts appropriately,
instead of needing the AF_* for the socket passed in.
# Something I'd like to write yet is proper signal() support. Unfortunately,
# the last time I tried, doing anything with the signal except ignoring it
# caused PHP to segfault. And to have decent socket support.. at least, from
# my experience, having signals is a good thing. Only problem.. to implement
# it, some changes would have to be made to the parser - since that is where
# the problems lie on handling asynchronous signals.
- Restore testing for Unicode charmap to what it was, it was working(?), this
shouldn't have any effect on other charmaps.
- Remove bogus message about not finding any Unicode mapping table.
- Complex fix for solving a problem with objects & method calls.
- Previous version is tagged PRE_METHOD_CALL_SEPERATE_FIX_PATCH.
- I need to check this fix on a server so if it doesn't work I will revert
- it.
- Still one misssing zend_list_addref() and small fix in pg_close.
Fixes #5349 (but not the way reporter wants...).
# There's something in the way database_connect/close functions behave when
# implemented the way as in mysql module that I find a bit strange... because
# of that extra zend_list_addref() there to prevent default link disappearing
# if it's not assigned to variable, when it is, I guess its' reference count
# is at this point 2, I can do: open a db link and assign it to a variable,
# close the link and still use it afterwards.
- Test pg_encoding_to_char instead (it's existence in libpq means that
PostgreSQL was compiled with multibyte support and client encoding
functions do something meaningful). Fixes #5417.
When using HTTP auth from PHP, fill in the %u custom log field so the
authenticated used id will get logged in the Apache access_log
@- When using HTTP auth from PHP, fill in the %u custom log field so the
@ authenticated used id will get logged in the Apache access_log (Rasmus)
projects / php / log