- removed not utilized user_allow_email column from schema
- removed inactive groups (they had no use at all, since inactive users are not able to login)
The only benefit those brought are distinguish users - but this is no longer needed too due to the inactive code present. This also allows us to retain group memberships as well as default settings for users being set inactive due to profile changes.
- rewrote user_active_flip to support multiple users and a mode, as well as coping with the aforementioned changes
- implemented updated jabber class to support SRV server records and for better jabberd2 support.
- jabber errors now logged to the error log with a full transaction
- fixed user_delete calls to include usernames where possible and also update last post information correctly
- implemented additioal checks to user management to cope with common mistakes
- On installation, guess the required mysql schema as best as possible. Users now only need to decide if they want to use the mysqli extension or not (mysqli selected by default) and no longer need to know their mysql version.
- founders do not need to re-activate their account on profile changes
- remove older session if re-authentication was successful (re-authentication always assigns a new session id)
- set the cookie directly instead of using php's function
- added inactive_remind to see which users got deactivated because of reminders (or re-activation) sent out
hopefully not introduced too many bugs - those testing with CVS releases, please concentrate on user registration, activation, profile changes (email/password)...
git-svn-id: file:///svn/phpbb/trunk@6436 89ea8834-ac86-4346-8a33-228a782c2dd0
- MySQL 3.x works now
- FirebirdSQL is now on the same level as MySQL and PostgreSQL, zero hacks exist inside the core code now
git-svn-id: file:///svn/phpbb/trunk@6422 89ea8834-ac86-4346-8a33-228a782c2dd0
- added optional MX and DNSBL checks
- added backtrace (triggering sql error) on error within sql_in_set as well as making sure it is handling an array
- let users having f_list access to a forum actually see the forum without a topic list and not displaying an error message - this allows for giving people access to subforums but not the parent forum without the need to add the (sub-)forum to the index.
- some additional bugfixes
git-svn-id: file:///svn/phpbb/trunk@6414 89ea8834-ac86-4346-8a33-228a782c2dd0
- MSSQL now works :P (sans UTF-8 support)
- The current schema files work well enough to install, etc. but must be tightened. Will get to it when I get some time...
git-svn-id: file:///svn/phpbb/trunk@6381 89ea8834-ac86-4346-8a33-228a782c2dd0
- take local server time into consideration for birthday/age calculation
- faster active topic search
- allow changing active topic time frame [Bug #4150]
- reload stylesheet on language change [Bug #4222]
git-svn-id: file:///svn/phpbb/trunk@6380 89ea8834-ac86-4346-8a33-228a782c2dd0
* Fix moderator roles to match their description
* Add new forum role for closer mapping of converted permissions
git-svn-id: file:///svn/phpbb/trunk@6375 89ea8834-ac86-4346-8a33-228a782c2dd0
- changed attachment handling a bit
- tried to remove target tags out of the code
- do not add session ids to urls for bots as well as not creating a new session on each page view for them
I bet i introduced some bugs too. ;)
git-svn-id: file:///svn/phpbb/trunk@6364 89ea8834-ac86-4346-8a33-228a782c2dd0
Note: This template variable is available but will not be used by default
git-svn-id: file:///svn/phpbb/trunk@6360 89ea8834-ac86-4346-8a33-228a782c2dd0
- adjusting acl_gets calls (they are or'd - make sure they get checked correctly based on the situation)
- automatically assign the u_download and u_search permission to the guest group
git-svn-id: file:///svn/phpbb/trunk@6351 89ea8834-ac86-4346-8a33-228a782c2dd0
- using E_USER_WARNING if an error occurred within the ACP (sadly not able to use it as a default for trigger_error - it seems to be hardcoded in PHP)
git-svn-id: file:///svn/phpbb/trunk@6320 89ea8834-ac86-4346-8a33-228a782c2dd0
- schema changes for first post and last post
- display in viewforum, search and ucp of first post
- update of database on posting
Still outstanding
- display of last post
git-svn-id: file:///svn/phpbb/trunk@6309 89ea8834-ac86-4346-8a33-228a782c2dd0
Add additional note about the install directory
Remove some empty tags from the template
git-svn-id: file:///svn/phpbb/trunk@6305 89ea8834-ac86-4346-8a33-228a782c2dd0
well, the super fast, ultra efficient, massively huge BBCode handling system was implemented differently on each DBMS. Although this provided the best performance, the solution was a bit hacky.
So what does this new thing do? We use base64 encoding to make everything nice and shiny, it turns into nice, safe characters that we can just jam into varchars on essentially any database. This has two implications: we must decode every bitfield we get AND we have slightly fewer IDs to work with. It goes down from 2040 BBCodes to 1512. We lose like a quarter of them :P
P.S. I hope nothing broke :P
git-svn-id: file:///svn/phpbb/trunk@6263 89ea8834-ac86-4346-8a33-228a782c2dd0
MSSQL
Firebird
SQLite
Oracle
Databases that lose cool points:
MySQL
PostgreSQL
MySQL recently-ish gained character set features. PostgreSQL has zero, none; one must create the database as unicode. The only way I can think of getting PostgreSQL to do what I want it to do would require something rather crazy...
git-svn-id: file:///svn/phpbb/trunk@6246 89ea8834-ac86-4346-8a33-228a782c2dd0
- made imageset naming more consistent
- updated every schema to be consistent and also fixed it (every db should install fine now)
-
git-svn-id: file:///svn/phpbb/trunk@6237 89ea8834-ac86-4346-8a33-228a782c2dd0
- shortening some db columns to meet the requirements
- correctly increase/decrease user post counts
- fix the topic title length bug(s)
git-svn-id: file:///svn/phpbb/trunk@6224 89ea8834-ac86-4346-8a33-228a782c2dd0
- added missing images to acp_styles [Bug #2623] and added a new image btn_warn
- use the style name when exporting a style instead of the non existant path [Bug #2343]
- fixed an incorrect call of acp_styles::install_style() [Bug #2325]
- always define $search_query in fulltext_mysql [Bug #3476]
git-svn-id: file:///svn/phpbb/trunk@6216 89ea8834-ac86-4346-8a33-228a782c2dd0
- Make MySQL schema a little different than the others...
- Fix MCP_TOPIC
git-svn-id: file:///svn/phpbb/trunk@6212 89ea8834-ac86-4346-8a33-228a782c2dd0
- removed the split_words array, introduced an enforced search_query
- the forum used for global topics in the search is now a forum, and no longer a category [Bug #2561]
- Bug #3404
- allow accessing reports by report_id, in contrast to mcp_queue this cannot just use the post id, since there can be multiple closed reports per post, so closed reports have to be accessed by report id, open reports, can optionally be accessed by report id or post id [Bug #3149]
- only attempt to unflag reported topics on closing a report when there are any without other reported posts [Bug #3057]
- updated fulltext_mysql to use the the search_query string
- overwrote the old fulltext_native with our improved version since it consumes too much time to maintain boths and we would switch to the improved version later anyway
- always show a link to search a user's posts even if the postcount is zero since he might only have posted in forums which do not count posts [Bug #3267]
git-svn-id: file:///svn/phpbb/trunk@6211 89ea8834-ac86-4346-8a33-228a782c2dd0
This commit should increase the total number of BBCodes from 31 to 2040. Some things to watch out for:
Each database likes to deal with binary data in its own, special way. They are, quite frankly, too cool for school.
MySQL, MSSQL and Oracle all allow me to send in a default value for their binary column using a hex number. However, MSSQL forces me to send the specific data as a hex number and thus we must CAST it.
PostgreSQL allows me to set a binary column, but with a twist. It demands that the default be in _octal_ and its datatype allows somewhere around a gigabyte's worth of BBCodes ( PGSQL users, we shut you down to 2040 for your own good! )
Firebird has no decent mechanism for allowing me to shuttle in binary data so I must force my way in. By virtue of triggers and a UDF, we ram in our default values.
SQLite is the most bizarre of them all. They have no mechanism for turning an ASCII code into a ASCII character. Because of this, we have a trigger and a UDF (just like Firebird!) but with a twist! The UDF is defined on the PHP side of things instead of SQL. SQLite also demands that it's data be encoded before being sent off.
Other notes:
- SQLite installs again :D
- Firebird nearly installs again :P
- Database backup is not screwed up :P
P.S.
I hope nothing broke :D
git-svn-id: file:///svn/phpbb/trunk@6209 89ea8834-ac86-4346-8a33-228a782c2dd0
- set ip_check to A.B.C. by default
- display postings in other encodings by default and present link to force the encoding as usual.
git-svn-id: file:///svn/phpbb/trunk@6198 89ea8834-ac86-4346-8a33-228a782c2dd0
comment -> attach_comment
new, forwarded, unread, marked, deleted -> pm_new, pm_forwarded, pm_unread, pm_marked, pm_deleted
module_name -> module_basename
value -> lang_value
- every column is now NOT NULL
- every column is now having a DEFAULT value
- hopefully mostly consistent across every db schema
- untested schemas: sqlite, oracle, firebird
git-svn-id: file:///svn/phpbb/trunk@6177 89ea8834-ac86-4346-8a33-228a782c2dd0
DOMAINs are simpler and thus the system is capable of determining what is _really_ going on (which is why it handles indexing for us :D)
git-svn-id: file:///svn/phpbb/trunk@6160 89ea8834-ac86-4346-8a33-228a782c2dd0
