1
0
mirror of https://github.com/DataTables/DataTables.git synced 2025-03-04 05:29:14 +01:00

120 Commits

Author SHA1 Message Date
Allan Jardine
4812d1d35b Dev: New: _DT_RowIndex is a private parameter that is attached to all TR elements in the table now. This allows fast reverse mapping from the row index to the aoData entry (typically this will be used for getting _aData). Before we would have to loop over all rows which could be very slow, but this reverse mapping allows several new options for optimisations and future enhancements.
Dev: Updated: All fnRender calls have been put into a single point - code reuse...
2012-01-20 21:48:45 +00:00
Allan Jardine
80472cdd13 Dev fix: Oops - inverted logic for column filter definitions 2012-01-15 10:45:28 +00:00
Allan Jardine
1e7f3066f9 Fix: A stored display start (iInitDisplayStart) of 0 would cause an undesired effect 2012-01-15 10:43:01 +00:00
Allan Jardine
91c63b8ad5 Fix: When a previous search column details is defined for column 0, it would fail incorrectly. Now fixed 2012-01-15 10:37:43 +00:00
Allan Jardine
e1146e2f9d Fix: _fnGetTdNodes would error if trying to get only the first row 2012-01-15 09:32:05 +00:00
Allan Jardine
dbf87ff5cd Performance: MUCH faster removing of DOM elements in the header and body of the table - 8060 - performance testing: http://jsperf.com/remove-detach/11 2012-01-14 09:48:52 +00:00
Allan Jardine
1077ebcf08 Performance: there can only be 1 open row per parent TR, so quit out the loop when found - 8060 2012-01-14 09:38:19 +00:00
Allan Jardine
91c8547f48 Fix: Couple of trivial optimisations for speed and size 2012-01-10 14:35:18 +00:00
Allan Jardine
7ae342dc7d Fix: Unable to set an sDom default if bJQueryUI:true is used 2012-01-07 16:58:58 +00:00
Allan Jardine
b689320911 Fix: In browsers that don't define JSON then there woudl be an error thrown 2012-01-07 13:31:57 +00:00
Allan Jardine
9809d71161 Dev update: Use $.extend to copy arrays and objects for the state savings, as we don't want plug-ins to be able to muck around with the DataTables internals by mistake (even with the slice, the objects in the arrays maintain their references) 2012-01-02 19:59:52 +00:00
Allan Jardine
f702ac5648 New: Allow state loading to be cancelled from the aoStateLoadParams callbacks by returning false (matches the capabilities of the old state loading methods) 2012-01-02 10:19:19 +00:00
Allan Jardine
1bdbe86da2 New: Init option - fnStateLoad.
New: Init option - fnStateLoadParams
New: Init option - fnStateLoaded
New: Init option - fnStateSave
New: Init option - fnStateSaveParams
New: Event - stateSaveParams
New: Event - stateLoadParams
New: Event - stateLoaded
Removed: Init option - fnStateLoadCallback
Removed: Init option - fnStateSaveCallback
Updated: Documentation and code tidy up of the updated state saving methods.
2012-01-02 10:12:39 +00:00
Allan Jardine
cf12ba5ab3 Update: Initial work on updating state saving code to use objects rather than constructing a JSON string. Note that this has involved removing fnStateLoadCallback and fnStateSaveCallback which are now replaced with a much more effective setup using fnStateSaveParams, fnStateLoadParams and fnStateLoaded (blog post to come about this). Code documentation and tidy up yet required 2012-01-02 09:21:19 +00:00
Allan Jardine
6a59b64a60 Fix: The settings finder from node should really use === rather than ==... 2011-12-28 11:16:34 +00:00
Allan Jardine
aeed78a08c Updated: Remove the sanity check for cols x rows === cells - we can't get to this code anymore due to the fnSetCellData check (that throws an error first), so this is dead code. 2011-12-28 11:14:53 +00:00
Allan Jardine
cb014e81d7 New: Add unit tests for fnCreatedCell and fnCreatedRow
New: New init option fnCreatedRow - very similar to fnCreatedCell but in this case used for TR elements
Updated: fnCreatedCell now also gets the column index passed in
2011-12-28 11:12:30 +00:00
Allan Jardine
4808f0edd7 Dev update: Remove the 'row' event. On experimenting with CellCreated I've found that I'm getting around 4'500 triggers / sec with Safari 5.1, which really isn't close to fast enough (about two orders of magnitude out from what I would like) which means that a table with 2000 rows and 9 columns would see an additional overhead of about 4 seconds on initialisation - ouch!!! As such I've removed the 'row' event here and it would seriously impact larger tables. events are still extremely useful, however, I would suggest that plug-ins should use the callback arrays rather than events since it is much faster. These arrays might be developed further into DataTables' own triggering methods in future. Something to to be aware of certainly going into 1.10. 2011-12-27 19:46:43 +00:00
Allan Jardine
5a2d4f8782 New: Add "processing" event 2011-12-27 18:37:28 +00:00
Allan Jardine
8c9f0b61e8 Updated: Nearly all callbacks now use an array and fire an event using the use DataTables standard method for dealing with callbacks. row, header, footer, init and pre-draw callback have been added to this new method. Only aoStateSave does not use method yet. fnCookieCallback and fnInfoCallback will not use the new methods.
Update: fnRowCallback - no need to return the TR element - it was a fairly pointless thing to do - if you want to replace the element then you need to do it with fnDrawCallback
2011-12-27 13:02:18 +00:00
Allan Jardine
ab9138b773 Updated: Initial pass at tidying up callbacks and events in DataTables - adding _fnCallbackReg and _fnCallbackFire to provide a single point for registering and firing callbacks in a unified way. Not all callbacks yet use these methods and some callbacks need to be converted to the array method (useful for plug-in developers) 2011-12-25 07:52:51 +00:00
Allan Jardine
94542c810a New: fnRender update - data object now also contains mDataProp along with the other information from before. There is also a second parameter passed to the function - the current value of the cell (was possible to get it from the data array, but this makes life a little easier if you want to get just that, which is the most common usage here). 2011-12-23 18:31:25 +00:00
Allan Jardine
0a1a2a17dc Updated: The call to _fnFilterCustom doesn't need to check if there are filters since the loop will effectively do that anyway. 2011-12-21 09:11:12 +00:00
Allan Jardine
74f3426284 Fix: Modify the filtering such that client-side filtering is not done when server-side processing is enabled - its just a waste of time since hte server will do it all - 7825 2011-12-21 09:07:44 +00:00
Allan Jardine
771df2ac2d Updates for documentation - document oApi as required 2011-12-15 13:54:23 +00:00
Allan Jardine
e8fb40640e Make warning message smaller - too much payload for something that is virtual never seen 2011-12-15 12:55:06 +00:00
Allan Jardine
65f34e23f3 Apply default column options - the column defaults weren't actually being included, nevermind applied. Now put them in DataTable.defaults.columns and have them work! 2011-12-15 12:38:08 +00:00
Allan Jardine
91858f9e86 Updated: Modification to the way table width is handled - if the 'width' attribute is on the table tag, then we use that to size the table. This makes having a table width of 100% much easier than before. The reason we need to do this is that there just doesn't appear to be a good way to get the relative width from stylesheets across browsers. 2011-12-15 09:43:31 +00:00
Allan Jardine
8714f7cc0e Fix: fnInfoCallback was not being executed in the DataTables' instance scope 2011-12-15 08:08:41 +00:00
Allan Jardine
e1d860205b Change the defaults object from attaching to DataTable.models to just the top level DataTable as DataTable.defaults - makes it much easier to access it when trying to set a default. 2011-12-14 16:22:58 +00:00
Allan Jardine
c44e10cf98 Tidy up the settings object - a lot of the settings should have null as the default since they are set by the initialisation.
Dev fix: Filtering usig anything other thant he API would always do case-sensitive matching
Doc fix: bAutoWidth wasn't documented
2011-12-14 16:02:03 +00:00
Allan Jardine
6db08eb107 New: Filtering now has a case insenstive option - this is bCaseInsensitive in oSearch, or the 6th parameter for fnFilter 2011-12-14 15:26:37 +00:00
Allan Jardine
b235c64c2a Language options always get extended with the initalisation object now - this allows you to use both a file source and the local script if you wanted. It also allows backwards compatiablity with language files that don't have the oAria namespace. 2011-12-14 13:39:06 +00:00
Allan Jardine
27488c4093 Tidy up use of typeof and undefined in DataTables - the main goal is to reduce code size under 70K here, but its a good opertunity to get this area sorted out. Note that the unit test update for fnInitComplete is because 'json' is now always passed through, although it will be undefined in anything but Ajax sourced client-side processing. 2011-12-14 13:35:49 +00:00
Allan Jardine
64a1b6b032 Copy the column options on top of the column settings object 2011-12-14 11:33:29 +00:00
Allan Jardine
b3e57df2ef Reduce code size slightly (not much in fairness - but every little helps...) 2011-12-14 11:27:45 +00:00
Allan Jardine
013eac9d0b Only the constructor should be referring to oJUIClasses - also in the constructor copy the class object rather than referencing 2011-12-14 11:15:17 +00:00
Allan Jardine
2bc7cd34c9 Remove redundant information in the message 2011-12-14 10:57:34 +00:00
Allan Jardine
169eb721d2 New: oLanguage.oAira.sSortAscending and oLanguage.oAria.sSortDescending language strings which are attached to the label for a column header when it can be sorted. Basically the same as before, but now customisable.
New: iTabIndex initialisaiton option - allow the developer to decide what tab index could be given to the table.
Reverted: Two button navigation using A tags again - important for if there is no CSS
New: ARIA supoprt for the full numebrs pagination style
Changed: Tidy up the way events are added to the pagination and headers to reduce code.
Changes: Complete ARIA support for column headers so we now consider bSortable
2011-12-14 10:38:20 +00:00
Allan Jardine
bec6085880 Update ARIA attributes based on feedback from Vinnie Young ( http://groups.google.com/group/free-aria/browse_thread/thread/96156d857f35e10d#msg_e451dd817dfe01b1 ). Label for column header now states the column title and the next sort. TBODY has a role of alert and TR elements in THEAD have a role of 'row' 2011-12-13 08:55:46 +00:00
Allan Jardine
3191ee098b Add the Ajax core JS file (forgot to commit it before!) 2011-12-13 08:18:42 +00:00
Allan Jardine
4e8f0fadd9 _fnArrayCmp is not used anywhere these days - ditch it 2011-12-12 09:38:34 +00:00
Allan Jardine
d7d3aeddea Fixes from JSHint and JSHint configuration. Note that I'm using a slightly modified JSHint to remove checking of mixed spaces/tabs and trailing white space due to issue 21 against JSHint (https://github.com/jshint/jshint/issues/21) - specifically it objects to my comment styling for the JSDoc comments, which I've got _everywhere_! 2011-12-12 07:43:54 +00:00
Allan Jardine
88b2990e53 Correction to the last commit - check against the default parameter rather than the default value! 2011-12-11 08:52:13 +00:00
Allan Jardine
9a94759fa4 Override the default sDom when using jQuery UI 2011-12-11 08:51:03 +00:00
Allan Jardine
d9e02ab5c8 Pull the Ajax functions into their own file 2011-12-11 08:44:28 +00:00
Allan Jardine
e5816e381c New: fnPageChange will now accept integers as well as the four current strings. The integer given represents the page that the table should jump to. Also update the internal calls to use this new ability as it is cleaner. 2011-12-11 08:41:55 +00:00
Allan Jardine
0d772b5eba Always add an ID to a table if it doesn't already have one
Add a class to the table to identify it as a DataTable
Update setAttribute to use id - works just as well and slightly smaller
2011-12-11 08:27:41 +00:00
Allan Jardine
d0458ca22c First effort at ARIA support - the table control widgets link themselves to the table now ('controls') and the column headers say what they do. The table is also described by the information control. Note that an ID on the table is needed for most of this, since ARIA can reiference elements. 2011-12-10 11:34:30 +00:00
Allan Jardine
d4366df398 Change how the state loading callback functions operate - previously it would require that a cookie be available before the state loading callback is called. Now is state loading occurs at all the callback function will be fired - allowing easer integration with server-side state storage. I don't actually see a downside to this and it should be perfectly backwards compatible. 2011-12-08 09:51:40 +00:00