summaryrefslogtreecommitdiff
path: root/src
Commit message (Collapse)AuthorAgeFilesLines
* typeck: When checking BOOL permit numbers because they have coercion rulesHEADmasterDaniel Silverstone2020-10-021-3/+7
| | | | Signed-off-by: Daniel Silverstone <dsilvers@digital-scurf.org>
* output_method_cdata: use output_ccode when possibleDaniel Silverstone2020-05-231-1/+1
| | | | Signed-off-by: Daniel Silverstone <dsilvers@digital-scurf.org>
* output: Use output_ccode for overloaded method bodiesDaniel Silverstone2020-05-231-3/+2
| | | | Signed-off-by: Daniel Silverstone <dsilvers@digital-scurf.org>
* parser: General AST which the IR recognises for overloadingDaniel Silverstone2020-05-231-13/+1
| | | | Signed-off-by: Daniel Silverstone <dsilvers@digital-scurf.org>
* duktape: Guess at and support more than pointer init argumentsDaniel Silverstone2020-05-222-3/+72
| | | | Signed-off-by: Daniel Silverstone <dsilvers@digital-scurf.org>
* add a mainpage to docs and outstanding todo entriesVincent Sanders2019-05-041-0/+10
|
* When operating on Window event attributes, pass NULL to dukkyDaniel Silverstone2019-05-041-3/+16
| | | | Signed-off-by: Daniel Silverstone <dsilvers@digital-scurf.org>
* Make file output use a common interfaceVincent Sanders2019-05-0310-574/+913
| | | | Signed-off-by: Vincent Sanders <vince@kyllikki.org>
* Generate all EventHandler kind outputsDaniel Silverstone2019-05-031-3/+6
| | | | Signed-off-by: Daniel Silverstone <dsilvers@digital-scurf.org>
* duk-libdom: Use PrimaryGlobal for Window to change method privateDaniel Silverstone2019-05-021-7/+16
| | | | | | | | | In order to support methods on `Window` being called via the global object (which does not set `this`) we notice that the interface we are generating the private data fetcher for as being the global one, and use `duk_push_global_object` rather than `duk_push_this`. Signed-off-by: Daniel Silverstone <dsilvers@digital-scurf.org>
* DUKKY: Switch to using NSLOG rather than old LOG or JS_LOGDaniel Silverstone2019-02-172-8/+8
| | | | | | | Also use the `dukky` logging category so that we have better separation of log messages regarding Duktape Signed-off-by: Daniel Silverstone <dsilvers@digital-scurf.org>
* Global variable errtxt is defined in two separate files.Samuel Holland2019-02-152-2/+2
| | | | This is a minor C standard conformance issue, and causes the build to fail with a duplicate symbol error when CFLAGS includes "-fno-common". The following patch fixes the issue by making the variables static.
* Fix YYLTYPE nonsense with bison >=3.0John-Mark Bell2018-01-201-0/+8
|
* remove the old JSAPI binding generation codeVincent Sanders2017-03-317-3760/+0
| | | | | | genbind has not generated valid JSAPI (spidermonkey) bindings in some time, remove the code to avoid confusion, its always available in the git history if necessary
* Operation generation: When argument type errors are detected, treat as error.tlsa/duktape2Michael Drake2017-03-221-3/+3
|
* Duktape 2.X: duk_error now returns a duk_ret_tMichael Drake2017-03-221-3/+3
| | | | | | | | | | | https://github.com/svaarala/duktape/blob/master/doc/release-notes-v2-0.rst#duk_error-duk_error_va-duk_throw-duk_fatal-have-a-return-value This change suppress warnings about ignoring the return value. However, it looks like we should be returning error in these cases rather than continuing. I'll do that in a separate commit.
* Duktape 2.x: duk_safe_call callbacks now have a void ptr user data param.Michael Drake2017-03-222-3/+3
| | | | https://github.com/svaarala/duktape/blob/master/doc/release-notes-v2-0.rst#duk_safe_call-userdata
* Align the private structs to quieten certain warnings from clangDaniel Silverstone2017-02-051-1/+1
|
* Ensure generated attributes register for correct eventsdsilvers/eventworkDaniel Silverstone2017-02-051-4/+7
|
* Make the #line a commentDaniel Silverstone2017-02-051-1/+1
|
* improve webidl AST dump outputVincent Sanders2016-11-271-9/+96
|
* remove unused node set APIVincent Sanders2016-11-272-8/+0
|
* restructure AST node creation to avoid castsVincent Sanders2016-11-276-69/+153
| | | | | | | | | | | | | | | | This changes Abstract Syntax Tree node creation for both webidl and genbind syntax tress. If a node is to be created with a numeric value instead of a pointer a separate API is now used instead of casting through void. This fixes parsing and AST building on big endian 64bit platforms where casting through void, which is completely undefined behaviour, generates different and non-functioning code. The solution in this patch is properly portable and correct without relying on casting at all. Thanks to James Clarke <jrtc27@jrtc27.com> for the original debugging and patch demonstrating how to work round the bug.
* use bounded string writingVincent Sanders2016-10-141-2/+2
|
* ensure webidl intercalate copes with empty webidlVincent Sanders2016-02-071-10/+14
|
* Bison after version 3.0 require different api prefix handlingVincent Sanders2016-02-071-2/+4
|
* fix ambiguity in grammar around class argumentsVincent Sanders2016-02-071-4/+2
|
* use unsued variable macro to remove warningVincent Sanders2016-02-071-2/+2
|
* Correctly annotate unused variables with a macroVincent Sanders2016-02-065-7/+18
|
* Add support for TreatNullAs=EmptyString extended attribute.John-Mark Bell2016-02-013-2/+20
|
* Expose nullable flag in IR and have generated getters use it.John-Mark Bell2016-02-013-2/+14
|
* Make generated string attribute getters cope with libdom returning NULL.John-Mark Bell2016-01-311-4/+8
| | | | Fixes #2415, #2418.
* simplify strndup detectionVincent Sanders2016-01-241-3/+1
|
* Improve strndup detectionVincent Sanders2016-01-241-1/+3
|
* Use attribute types from IR for generated event handlersVincent Sanders2015-11-052-20/+149
|
* extend IR to have a separate map for attribute typesVincent Sanders2015-11-052-32/+92
|
* Add generated code for short integer valuesVincent Sanders2015-11-011-0/+53
|
* split heuristic code generation out from interface logicVincent Sanders2015-11-014-185/+230
|
* extend the interface attribute setter automatic code generation to cover longsVincent Sanders2015-10-251-0/+29
|
* improve unimplemented warning to include type modifierVincent Sanders2015-10-243-7/+16
|
* extend the interface attribute getter automatic code generation to cover longsVincent Sanders2015-10-241-0/+24
|
* Extend IR attribute entries to contain the type modifilerVincent Sanders2015-10-242-17/+43
|
* Fix putforwards generation to cope with getters leaving a dirty stackVincent Sanders2015-10-141-5/+4
| | | | Kinnison did all teh heavy tjinking on this!
* slightly modify the IDL to c name translationVincent Sanders2015-10-131-8/+14
| | | | | | | This tweaks the IDL name to c name translation to not leave isolated characters. e.g. HTMLIFrameElement became html_i_frame_element now its html_iframe_element HTMLDList became html_d_list and is now html_dlist
* Extend the permitted values of the binding method namesVincent Sanders2015-10-121-2/+21
|
* show the property type on unimplemented getter warningVincent Sanders2015-10-121-2/+4
|
* show the property type on unimplemented setter warningVincent Sanders2015-10-093-2/+60
|
* Improve the generated instanceof in the binding to take an indexVincent Sanders2015-10-091-16/+16
|
* Add automatic generation of property getters and settersVincent Sanders2015-10-083-50/+241
| | | | | | | | | | | This allows the binding to omit specifying code for class property getters and setters. The omitted code will be generated by genbind to directly call suitable libdom accessors. The type of the property (string, boolean etc.) is derived from the IDL. If this type is incorrect, or is a type the generator cannot automatically produce, then the property will be treated like it has no binding implementation and generate unimplemented warnings.
* Split out idl to c name conversionVincent Sanders2015-10-083-64/+69
|