Jon Parise [Fri, 6 Apr 2001 16:24:55 +0000 (16:24 +0000)]
Comment out the 'include_path' variable by default.
# Leaving it blank overrides the default include path, which contains the
# PEAR directory.
# The other option is to have the build system generate this file and expand
# @PEAR_INSTALLDIR@ on this line to reflect the system default.
- Lock the whole thing for now in order not to have a race condition.
It shouldn't effect performance too much because usually threads are
reused a few times before they are nuked by the web server, at least
under IIS and Zeus.
- Revert previous fix because dtor's which access thread local storage
might stop working. Need to fix this! We can either mutex the whole thing
which is slow (this is how it was before the race condition was introduced)
or find a different solution.
Jason Greene [Wed, 4 Apr 2001 04:41:01 +0000 (04:41 +0000)]
This is a temporary fix to the previous patch that breaks the build.
This keeps the functionality of searching for glibtool or libtool,
but have you checked the rest of the build process to see if this is
compatible? Regardless, 'which' returns a value on some os's when nothing is
found which breaks the previous patches test. Also the unary negation after
an if is not compatible with true bourne.
changed the flags parameters to be a bitfield + constants (I don't know what
I was thinking!). Added a couple of extra parameters so that the equivalent
functionality of the openssl smime command can be pretty much used in full.
Added some more error reporting.
Nick Gorham [Mon, 2 Apr 2001 15:18:47 +0000 (15:18 +0000)]
Added small change to php_odbc module, to check for failed SQLDisconnects
and to close any outstanding transactions if the call fails, then disconnect
again. This can cause chaos with SQL Server, this regards a SELECT as
starting a transaction, and will leave it open if the result set is not all
read.
Ulf Wendel [Thu, 29 Mar 2001 15:01:16 +0000 (15:01 +0000)]
- package XML for the Cache
I'm not sure how to use the package.dtd to document packages and if at all the Cache can be it's own package, but one has to start with it... Sterling, Stig can you give me a hint?
Tomas V.V.Cox [Wed, 28 Mar 2001 23:52:21 +0000 (23:52 +0000)]
pgsql.php
* better handling of $rownum param in fetchInto()
* changed property numrows[] to num_rows[]
DB.php
* $rowmun in fetch* now defaults to NULL
mysql.php
* added default $fetchmode to ordered
* removed extra checks in connect()
* fetchrow() now uses fetchInto()
* added the "fetch absolute row numbers" feature to fetchInto()
(not tested)
ifx.php
* better handling of $rownum param in fetchInto()
Dan Kalowsky [Tue, 27 Mar 2001 15:46:27 +0000 (15:46 +0000)]
The last fix was missing a &, Marc Boeren either made the patch wrong, or I applied it improperly. Anyways, Sebastian Bergmann discovered it, and supposedly is fixed with this.
Tomas V.V.Cox [Mon, 26 Mar 2001 23:31:49 +0000 (23:31 +0000)]
pgsql.php
* connect() always use pg_connect string instead of the deprecated params
mode
* removed duplicated functions prepare() and execute() (now in common.php)
* pgsqlRaiseError() always fills native error param on DB_error objs
* added third param $rownum to fetchInto() so users can fetch also absolute
row numbers
* changed fetchRow() to use fetchInto() (can not erase, still used in get*()
from common.php
DB.php
* added third param $rownum to fetchInto()/fetchRow() so users can fetch
also absolute row numbers
* changed fetchRow() to use fetchInto()
projects / php / log