* Change archiving config page layout
I've changed some wording and moved actions into a
maintenance section.
* Update purge action
Now we have more control on the purge action. The configuration allows
us to choose what to keep and what to discard in a more precise way.
At the moment, the configuration applies for all feeds.
* Add purge configuration on feed level
Now the extend purge configuration is available on feed level.
It is stored as attributes and will be used in the purge action.
* Update purge action
Now the purge action uses the feed configuration if it exists and
defaults on user configuration if not.
* Add empty option in period list
* Fix configuration warnings
* Add archiving configuration on categories
See #2369
* Add user info back
* Add explanations in UI
* Fixes for SQLite + error + misc.
* Fix invalid feed reference
* Short array syntax
Only for new code, so far
* Fix prefix error
* Query performance, default values
Work in progress
* Fix default values and confirm before leaving
Form cancel and confirm changes before leaving were broken.
And start taking advantage of the short echo syntax `<?= ?>` as we have
moved to PHP 5.4+
* More work
* Tuning SQL
* Fix MariaDB + performance issue
* SQL performance
* Fix SQLite bug
* Fix some attributes JSON encoding bugs
Especially for SQLite export/import
* More uniform, fix bugs
More uniform between global, category, feed settings
* Drop special cases for old articles during refresh
Instead will use lastSeen date with the new archiving logic.
This was generating problems anyway
https://github.com/FreshRSS/FreshRSS/issues/2154
* Draft drop index keep_history
Not needed anymore
* MySQL typo
Now properly tested with MySQL, PostgreSQL, SQLite
* More work for legacy values
Important to avoid overriding user's preference and risking deleting
data erroneously
* Fix PHP 7.3 / 7.4 warnings
@aledeg "Trying to use values of type null, bool, int, float or resource
as an array (such as $null["key"]) will now generate a notice. "
https://php.net/migration74.incompatible
* Reintroduce min articles and take care of legacy parameters
* A few changes forgotten
* Draft of migration + DROP of feed.keep_history
* Fix several errors
And give up using const for SQL to allow multiple database types (and we
cannot redefine a const)
* Add keep_min to categories + factorise archiving logic
* Legacy fix
* Fix bug yield from
* Minor: Use JSON_UNESCAPED_SLASHE for attributes
And make more uniform
* Fix sign and missing variable
* Fine tune the logic
* First draft of custom tags
https://github.com/FreshRSS/FreshRSS/issues/928https://github.com/FreshRSS/FreshRSS/issues/1367
* SMALLINT to BIGINT for id_entry
And uppercase SQL types
* Fix layout for unreads
* Start UI menu
* Change menu order
* Clean database helpers
https://github.com/FreshRSS/FreshRSS/pull/2027#discussion_r217971535
* Travis rules do not understand PostgreSQL constants
Grrr
* Tag controller + UI
* Add column attributes to tags
* Use only favicon for now, for label
* Fix styling for different themes
* Constant for maximum InnoDB index length in Unicode
https://github.com/FreshRSS/FreshRSS/pull/2027#discussion_r219052200
(I would have personnally prefered keeping the readability of a real
value instead of a constant, in this case of many SQL fields)
* Use FreshRSS_Factory::createCategoryDao
* Add view of all articles containing any tag
* Fix search in tags
* Mark as read tags
* Partial auto-update unread tags
* More auto update tag unreads
* Add tag deletion
* Do not purge tagged articles
* Minor comment
* Fix SQLite and UI bug
* Google Reader API support for user tags
Add SQL check that tag names must be distinct from category names
* whitespace
* Add missing API for EasyRSS
* Compatibility SQLite
Problematic parentheses
* Add SQL DISTINCT for cases with multiple tags
* Fix for PostgreSQL
PostgreSQL needs some additional type hint to avoid "could not determine
data type of parameter $1"
http://www.postgresql-archive.org/Could-not-determine-data-type-of-parameter-1-tp2171092p2171094.html
* [ci] Add Travis
* Exclude some libs
* Semi-auto whitespace fixes
* line length in SQLite
* Exclude tests from line length
* Feed.php line length
* Feed.php: get rid of unnecessary concat
* Feed.php: line length
* bootstrap.php: no newline at end of file
* Allow concatenating across multiple lines
* Add Travis badge
* do-install line length
* update-or-create-user line length
* cli/create-user line length
* tests/app/Models/SearchTest.php fix indentation
* tests/app/Models/UserQueryTest.php fix indentation
* tests/app/Models/CategoryTest.php fix indentation
* [fix] PHP 5.3 on precise
* cli/do-install no spaces
* cli/list-users line length
* cli/reconfigure line length
* empty catch statements
* api/index line length nonsense
* spaces before semicolon
* app/Models/EntryDAO bunch of indentation
* extra blank lines
* spaces before comma in function call
* testing tabwidth
* increase to 10
* comment out tabwidth line
* try older phpcs version 3.0.0RC4
* line length exception for app/install.php
* proper spaces
* stray spaces in i18n
* Minz/ModelPdo line length
* Minz whitespace
* greader line length
* greader elseif placement
* app/Models/Feed.php spacing in function argument
* ignore php 5.3
* app/Models/ConfigurationSetter.php stray whitespace
* EntryDAOSQLite line length
* I vote for higher max line length =P
* ignore SQL
* remove classname complaint
* line length/more legible SQL
* ignore line length nonsense
* greader line length
* feedController issues
* uppercase TRUE, FALSE, NULL
* revert
* importExportController lowercase null
* Share.php default value not necessary because ! is_array () a few lines down
* CategoryDAO constants should be UPPERCASE
* EntryDAO reduce line length
* contentious autofix
* Allow failures on all versions of PHP except 7.1 because reasons
It is now possible to combine multiple keywords to do a search. The separation of concern is better now since the search extraction is not done in the DAO anymore.
At the moment, a multiple keyword search is stored as this. It could be nice to have it rendered differently in the search page to make it more readable.
At the moment, there is a problem with search enclosed by ". Same search works well when enclosed by '.
I figured that the code for the search could be extracted from the context to have separation of concern.
It supports multiple keywords. It suports also multiple tag keywords.
This feature is not in use at the moment, but it will be handy to reorganize the query building process. It allows to have more than one keyword in the search box.
Full tests are available as well.
It probably needs a refactoring later, but I think this is the first step to make the application full object oriented and testable.
- FreshRSS_Context::$conf is replaced by FreshRSS_Context::$user_conf
- Introduce FreshRSS_Context::$system_conf
- Remove FreshRSS_Configuration object
See https://github.com/FreshRSS/FreshRSS/issues/730
- Use only Minz_Configuration
- register() method to load a new configuration file
- get() to get a configuration
- new exceptions related to configuration
- fix a list configuration calls to have FRSS working
Current problems to resolve:
- How to handle configuration param verifications (i.e. check auth_type
is a value from none, http_auth, persona or form)
- We must use $conf = Minz_Configuration::get('system'); $general_conf = $conf->general;
to access global system configuration which is quite annoying. How to change that?
See https://github.com/FreshRSS/FreshRSS/issues/730
Before, when the article while marked as read while scrolling and auto removed after reading, the display was in the middle of the following article.
Now, the article is forced to be a sticky article so the user display is forced to be on top of the following article.
This is a new reading option to hide articles when they are read. The hidding process occurs when the article is left for an other article. This way, even when the article is marked as read on opening, it is hidden only while navigating to an other article.
I'm not really happy with the behavior when the "mark while scrolling" option is enabled. Please review.
It is missing the i18n since we're not supposed to push them before it exists on i18n.freshrss.org. Or maybe I misunderstood the process.
See #476
- Seperate normal, global and rss outputs in dedicated actions (NOT WORKING YET!)
- Rewrite aside_flux and nav_menu to use Context object
- Improve Context object
See https://github.com/marienfressinaud/FreshRSS/issues/634