summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* keyboard events: Fix get location casting wrapper.HEADmasterMichael Drake6 days1-2/+2
|
* keyboard events: Update to latest DOM spec.Michael Drake6 days3-83/+200
|
* ui events: constructor: initialise the detail attribute.Michael Drake6 days1-2/+1
|
* keyboard events: constructor: initialise the key location.Michael Drake6 days1-4/+1
|
* keyboard events: Expose the creation function.Michael Drake6 days2-3/+4
|
* namespace: Support finalising the namespace stringsDaniel Silverstone2019-09-072-3/+7
| | | | Signed-off-by: Daniel Silverstone <dsilvers@digital-scurf.org>
* element: Issue DOMSubtreeModified *after* updating attr valueDaniel Silverstone2019-08-041-4/+4
| | | | Signed-off-by: Daniel Silverstone <dsilvers@digital-scurf.org>
* html_text_area_element: Set node text content on value changeDaniel Silverstone2019-08-041-1/+1
| | | | Signed-off-by: Daniel Silverstone <dsilvers@digital-scurf.org>
* Update component version for releaserelease/0.4.0Vincent Sanders2019-07-171-1/+1
|
* String: remove innards from public header.John-Mark Bell2019-06-292-14/+14
| | | | | | | | | A dom_string is the first member of the underlying dom_string_internal and thus their addresses are equivalent. We shouldn't be exposing the internal alignment requirements through the public header at all, so remove the bogus _ALIGNED and cast through void internally to tell the compiler that we know what we're doing.
* Update Component version for releaserelease/0.3.4Vincent Sanders2019-06-261-1/+1
|
* node.c: _dom_node_set_text_content() fix leaksDaniel Silverstone2019-06-031-4/+9
| | | | | | | There were a number of text node leaks in the function to set a node's text content. This corrects those. Signed-off-by: Daniel Silverstone <dsilvers@digital-scurf.org>
* Assert instead of checkingDaniel Silverstone2019-05-061-1/+4
| | | | | | | At this point, we MUST have an element created, so assert it rather than half-heartedly checking. Signed-off-by: Daniel Silverstone <dsilvers@digital-scurf.org>
* Add dom_node_contains()Daniel Silverstone2019-05-042-0/+40
| | | | Signed-off-by: Daniel Silverstone <dsilvers@digital-scurf.org>
* Begin to support SCRIPT elements flagsDaniel Silverstone2019-05-044-0/+91
| | | | Signed-off-by: Daniel Silverstone <dsilvers@digital-scurf.org>
* 99.9% of all events are trusted, so default to trustedDaniel Silverstone2019-05-031-1/+1
| | | | Signed-off-by: Daniel Silverstone <dsilvers@digital-scurf.org>
* Include: Switch to using stdint.h rather than inttypes.hDaniel Silverstone2019-02-161-1/+1
| | | | | | | | We don't need the formatting strings, so instead include just the types themselves. This has the beneficial effect of making libdom compilable under WIN32. Signed-off-by: Daniel Silverstone <dsilvers@digital-scurf.org>
* DOMTSHandler: Unref nodes in for-each blocksDaniel Silverstone2019-02-161-2/+16
| | | | | | | In order to satisfy (hopefully) the last of the sanitize checks, ensure that we unref nodes during `for-each` iterations. Signed-off-by: Daniel Silverstone <dsilvers@digital-scurf.org>
* Fix libdom departures from "always return a ref" in public API.Michael Drake2019-02-164-4/+3
| | | | | | | | | | | | | | | | | | | The DOM test code expects nodes returned by libdom APIs to always be reffed. So this change refs them. The remaining leaks detected by address sanitiser in the dom test suite are a problem with the DOM test generator: * For test_level1_html_HTMLSelectElement08, the problem is that in the while loop, it doesn't unref `voption` at the end of the loop, in the test code. * For test_level1_html_HTMLTableElement07, the problem is the while loop doesn't unref `vrow` at the end of the loop in the test code. * For test_level1_html_HTMLTableElement09, it's the same for variable, `vbodies`.
* Update component version for releaserelease/0.3.3Vincent Sanders2018-08-221-1/+1
|
* Attempt to fix m68k cross-compile warnings.Michael Drake2018-08-071-2/+2
|
* Attempt to fix m68k cross-compile warnings.Michael Drake2018-08-071-11/+10
|
* Attempt to squash m68k cross-compile warnings.Michael Drake2018-08-071-2/+2
|
* Attempt to appease m68k CI cross-compiler.Michael Drake2018-08-071-5/+1
|
* DOMTSHandler: Ensure HTMLOptionsCollection is unreffed.Michael Drake2018-08-031-2/+3
|
* HTMLTableRow.insertCell: Fix refcounting and error paths.Michael Drake2018-08-031-13/+16
|
* HTMLTableElement.deleteRow: Squash leaks and fix error paths.Michael Drake2018-08-031-20/+25
|
* HTMLDocument: Squash leaks of document.documentElement.Michael Drake2018-08-031-4/+12
|
* HTMLDocument.body: Squash leak of node list.Michael Drake2018-08-031-0/+1
|
* HTMLCollection.namedItem: Squash leak of id_name.Michael Drake2018-08-031-15/+17
|
* dom_html_table_row_element_get_row_index: Squash t_head node leak.Michael Drake2018-08-031-5/+3
|
* HTMLTableElement.insertRow: Fix lifetimes and error handling.Michael Drake2018-08-032-53/+75
|
* HTMLTableSectionElement.insertRow: Fix lifetimes and error handling.Michael Drake2018-08-031-25/+25
|
* Cleanups in the DOMTS should be safeDaniel Silverstone2018-07-291-1/+1
|
* More cleanup for iterating in test suiteDaniel Silverstone2018-07-292-1/+7
|
* Various extra cleanups in list implementation for test suiteDaniel Silverstone2018-07-291-1/+5
|
* Make API note that unreffing is up to the caller for dom_nodelist_item()Daniel Silverstone2018-07-291-0/+3
|
* Unref nodes properly in foreachDaniel Silverstone2018-07-291-0/+7
|
* Clean up nodes and listsDaniel Silverstone2018-07-291-0/+8
|
* Update libdom python gdb to python 3Daniel Silverstone2018-07-291-4/+4
|
* Ensure we clean up if load_html fails in the test runnerDaniel Silverstone2018-07-291-1/+11
|
* Correct COLLECT logic in walk_logic_adjacent_text_in_order to stop leakDaniel Silverstone2018-07-291-8/+19
|
* README: Document testsuite requirement.Michael Drake2018-07-281-0/+1
|
* update component version for 0.3.2 releaserelease/0.3.2Vincent Sanders2017-10-131-1/+1
|
* Unregister all instances of a listenerDaniel Silverstone2017-02-041-3/+12
| | | | | | | Sometimes we register a listener a lot on an event target. This change allows us to pass NULL as the event type to the remove call in order to remove every instance of the given listener, regardless of event type and capture vs. bubble.
* Change dom node type count to last value plus one.Michael Drake2017-02-042-1/+2
|
* Ignore http and https addresses when fetching entity refsChris Young2016-11-201-0/+13
| | | | | The fopen() causes a requester on AmigaOS which I'm bored of hearing about. See also bug 2313
* Add getter/setter for event_target's is_trustedDaniel Silverstone2016-11-193-0/+38
|
* Add an is_initialised flag to eventsDaniel Silverstone2016-11-192-0/+18
|
* Add support to retrieve if an event is in dispatchDaniel Silverstone2016-11-192-0/+21
|