Revisions for framework/Prefs/Prefs/sql.php

Happy New Year

Tags: HORDE_3_3, HORDE_3_3_RC1, HORDE_3_2_2, HORDE_3_2_1, HORDE_3_2, HORDE_3_2_RC4, HORDE_3_2_RC3, HORDE_3_2_RC2

Happy New Year

Merge from HEAD.

Jon has assigned his copyrights to The Horde Project.

Tags: BEFORE_MERGE_3_2, HORDE_3_2_RC1, HORDE_3_2_ALPHA

Bump year, looks like a missed a few last time.

Typo.

Existence check.  Fixes a possible fatal error when configuring for the first time and going from no pref backend to a sql backend.

Happy New Year

Tags: HORDE_3_1_9, HORDE_3_1_8, HORDE_3_1_7, HORDE_3_1_6, HORDE_3_1_5, HORDE_3_1_4, HORDE_3_1_4_RC1

Fix two shared preference bugs, both of which could cause Bug #4378:
- load the defaults for shared preferences into the horde scope, not the scope
of the app that marks them shared
- retrieve shared prefs when loading the horde scope even though we might not
have a conf for them then, and don't overwrite the retrieved value when we
_do_ load the conf later.

[cjh] Rewrite the Prefs code to keep the same Prefs object through multiple
scope changes, instead of re-loading the Horde prefs, or prefs for
a scope that had already been loaded. Fixes bugs when making changes
to preferences and then switching scopes again (Bug #2838).

This is a big change to the prefs system, and while I've tested all of
it as best as I can, and the SQL driver I'm pretty confident in, this
needs testing before we even think about merging it. It would all be
BC, though.

Please let me know (or update the bug) if this works or doesn't work
for you, especially people using lots of hooks and other more edge
cases, and also let me know if you think it should be merged or not.

CS - use line continuations instead of .=

log more error information.

standard reference style

One more query to log.

Condense redundant code, log query.

MFH: Horde_String not used in all prefs backends.

Tags: HORDE_3_1_3, HORDE_3_1_2

Horde_String is not required by all prefs backends.

phpdoc cleanups
cache cleanups.  cacheUpdate() is now only called with the preferences that
need to be updated, instead of always updating the entire prefs array. On
my server, this reduced the number of assignments in cacheUpdate() from 2173
to 379. i think there is still some more tweaking that can be done in here.

Bug: 3372
MFH: Prefs.php - 1.161-1.163; imsp.php - 1.12; ldap.php - 1.105;
sql.php - 1.114

Tags: HORDE_3_1_1, HORDE_3_1

Don't think we need the '_filled' entries since we can simply check for the
presence of the 'scopename' in the prefs_cache array.

Warn user if loading prefs fails.

MFH 1.112

Tags: HORDE_3_1_RC3, HORDE_3_1_RC2

MFH

Typo.

use DB_PORTABILITY_RTRIM for mssql connections

Happy new year

Tags: HORDE_3_1_RC1

Happy new year\!

- Don't call Horde::fatal() in prefs drivers
- Remove call to Horde::serviceDegraded()
- Go back to only connecting to the backend if we actually need to - most
page loads should never hit the prefs backend due to the session cache.

Updates for SQL split servers
- use the seperate server parameters, they were being ignored before
- add split support to the History API
- default to the write server, not the read server
- create the _write_db connection in the SQL sessionhandler

MFH 1.107

Work around bug in some php versions with shutdown functions.
Bug: #3016

add documentation about the read/write params required

split read/writes, Token_sql was missing var and disconnect

remove the failover stuff from framework_3

MFH: failover and some typo fixing

Add failover support to Prefs too.
Fix what seems some typos in imsp ($this->_params).

Merged from HEAD.

MFH: Remove "@access public" and fix wrapping.

Tags: HORDE_3_0_12, HORDE_3_0_11, HORDE_3_0_10, HORDE_3_0_9, HORDE_3_0_8, HORDE_3_0_7, HORDE_3_0_6, HORDE_3_0_6_RC1, HORDE_3_0_5, HORDE_3_0_5_RC2, HORDE_3_0_5_RC1

Remove "@access public" and fix wrapping.

Tags: BEFORE_MERGE_3_1

MFH: Remove object keywords from phpdoc.

Remove object keywords from phpdoc.

MFH: 1.101

Don't convert booleans.

MFH 1.100

Cache converted values.

MFH: "optional" not needed in phpdoc.

"optional" not needed in phpdoc.

MFH.

MFH.

Clean up required/optional parameters in phpdoc.

This patch removes slow array_push() calls where possible and uses the query()
function instead of seperate prepare()/execute() calls.
Bug: 1706
Submitted by: selsky (at) columbia (dot) edu

MFH: Don't require 'database'

Don't require 'database'.

MFH: make these drivers work with SQLite

more docs/comment tweaking

Tweak required parameters so that these drivers work with SQLite.

MFH: bind variable usage in sql drivers (selsky@columbia.edu, Bugs
#1651, #1652, #1653).

[cjh] Use bind variables in the Prefs and Token SQL drivers
(selsky@columbia.edu, Bugs #1652, #1653).

Happy New Year

Happy New Year

Tags: HORDE_3_0_4, HORDE_3_0_4_RC2, HORDE_3_0_4_RC1, HORDE_3_0_3, HORDE_3_0_3_RC1, HORDE_3_0_2, HORDE_3_0_1

Use new DB 1.6 portability settings instead of old "optimize" "portability".

Tags: FRAMEWORK_3, HORDE_3_0, HORDE_3_0_RC3

Don't convert charset of prefs when storing them into the backend but keep
them already converted in the prefs cache.
This is necessary for example for serialized data like identities. When we
retrieved such prefs with another charset than we used to store them, the
conversion happened on the serialized string, breaking the unserialization
because the string lengths were not correct anymore.
Code that needs to store serialized data can now do the conversion itself,
serialize the data and then store it while telling the prefs backend to *not*
convert it automatically.

comment wrapping

Use prepare/execute for pref values for better support of large fields
on some database backends.

Bug: 443
Submitted by: selsky@columbia.edu

Add clear() method. Someone has to implement that for the LDAP driver.

line up

Fix some charset conversion weirdness ("W. Tasin" <tasin@fhm.edu>).

Tags: HORDE_3_0_ALPHA

Make sure to pass DB a password, even if it's empty.

[cjh] Prefs::getPref() has been deprecated and is no longer present.

Make sure the default preferences are always loaded, even if not using any
preference backend.

revert

this allows an pref value which is locked subsequently in the prefs.php config file to override also the value stored by the user in the backend previously, any problems with doing this?

Happy new year everyone.

fix log level for this query (Daniel Eckl <daniel.eckl@gmx.de> ).

package name

this is only shared if the scope is 'horde'

Here's another crack at the preferences problem. The logic for retrieve() seems to
be working correctly, so now we call setDefaults() only if we're going to actually
hit the backend in retrieve(), and don't call setDefaults() explicitly anywhere else.

This may mean a few things need to change for guest preferences, but we'll sort
that out.

Missed that one.

No gettext in logs please.

Add _convertFromDriver() and _convertToDriver() utility methods for charset
conversions.

Try to sort out the prefs caching logic.

More simplification of preferences code. listAll() isn't needed anymore.

Clean up preferences a bit:
- setDefaults() is now a private method (_setDefaults()), and only called by
the Prefs constructor.
- Everything we need gets stored in the session prefs cache. Therefore, we
only need to read in the prefs.php files once per session.
- clean up retrieve() a bit; we *never* pass it a list of explicit preferences, so
simplify by always loading all prefs per current scope.

Use Horde::assertDriverConfig().

More doc improvements

LOG_* constants -> PEAR_LOG_*

Don't mark newly loaded preferences as dirty.

Submitted by: Jeff Seifert <seifert@everybody.org>

Honor db persistent connections parameter in all drivers.

bloody typos...

MFH 1.62

Bug: 1224

Tags: HORDE_2_2_4_RC2, HORDE_2_2_4_RC1, HORDE_2_2_4

Cast pref_value to string so it doesn't quote a boolean as a boolean and
make Postgresql (and presumably others) complain.

Bug: 1224

MFH 1.60

typos.

Scope was not being set correctly in setPref().
When non-horde scope passed in, make sure that we update _that_ scope
and not a variable with the same name under the horde scope.
documentation.

minor cleanups.

style, is_a().

Eliminate some array_key_exists() calls.
style.

Add getCharset() to return the backend's charset.

Check for required 'charset' parameter.

Convert charset.

Add and fix @since tags.

Bump year.

Tags: HORDE_2_2_3, HORDE_2_2_2, HORDE_2_2_1, HORDE_2_2_1_RC1, HORDE_2_2, HORDE_2_2_RC2, HORDE_2_2_RC1