summaryrefslogtreecommitdiff
path: root/fxjs
AgeCommit message (Collapse)Author
2018-02-15Make the CFX_XMLNode a MaybeOwned pointerDan Sinclair
This CL removes the HasOwnXML flag from the CXFA_Node objects and instead uses a MaybeOwned pointer to keep track of the XML nodes. Change-Id: Ie678258247ec21ecb15c639647b189e140586d25 Reviewed-on: https://pdfium-review.googlesource.com/26811 Reviewed-by: Tom Sepez <tsepez@chromium.org> Commit-Queue: dsinclair <dsinclair@chromium.org>
2018-02-14Fix lifetime probe issue in CJX_Objectchromium/3348Dan Sinclair
This CL removes the UnownedPtr to the CXFA_LayoutItem from CJX_Object. This is because the CJX_Object will be destroyed by the CXFA_Node which is destroyed in the CXFA_Document destructor (due to the vector of unique_ptr being destroyed). The CXFA_LayoutItem will be freed in the LayoutProcessor which also lives in the CXFA_Document. Bug: chromium:807215 Change-Id: I86040e154ee2e5d461fc4d3565a10a9181680207 Reviewed-on: https://pdfium-review.googlesource.com/26851 Reviewed-by: Tom Sepez <tsepez@chromium.org> Commit-Queue: dsinclair <dsinclair@chromium.org>
2018-02-14Cleanup CXFA_Node::SetFlagDan Sinclair
The SetFlag method has a signature of (flag, bool) which, looking at the callsites makes you think it will set the flag to the boolean. This isn't what happens, the flag is always set to true and the boolean decides if we execute some notification code. This CL splits SetFlag into SetFlag(flag) and SetFlagAndNotify(flag) to make it a lot clearer what is happening. Change-Id: I3e067c87532cce10b94bda10cc88feb62b948eb0 Reviewed-on: https://pdfium-review.googlesource.com/26750 Reviewed-by: Ryan Harrison <rharrison@chromium.org> Commit-Queue: dsinclair <dsinclair@chromium.org>
2018-02-14Cleanup CXFA_Document const-nessDan Sinclair
This CL marks methods const where possible and removes redundant methods from CXFA_Document. Change-Id: I34804425469dfd52469daba68f64a4b511e5d4d0 Reviewed-on: https://pdfium-review.googlesource.com/26730 Commit-Queue: Ryan Harrison <rharrison@chromium.org> Reviewed-by: Ryan Harrison <rharrison@chromium.org>
2018-02-13Split CFX_XMLNode::SaveXMLNode apartDan Sinclair
This CL removes the switch from SaveXMLNode and moves the required code into override methods in the child classes. The method is renamed from SaveXMLNode to just Save. Change-Id: I2011b80525e99635c573b4e0cf977e94f6b7cea6 Reviewed-on: https://pdfium-review.googlesource.com/26590 Reviewed-by: Tom Sepez <tsepez@chromium.org> Commit-Queue: dsinclair <dsinclair@chromium.org>
2018-02-13Cleanup CFX_XMLNode pointersDan Sinclair
This CL cleans up hte CFX_XMLNode pointers. Each pointer has been renamed to make the usage clearer, the NodeItems method has been removed in favour of distinct accessors and the node pointers have been made private. Change-Id: I5459a77a0ae93b08741a0cd59266ef9c81ddad75 Reviewed-on: https://pdfium-review.googlesource.com/26550 Commit-Queue: dsinclair <dsinclair@chromium.org> Reviewed-by: Tom Sepez <tsepez@chromium.org>
2018-02-13Cleanup CFX_XMLNode and friendschromium/3347Dan Sinclair
This CL removes unused methods from CFX_XMLNode, adds an AppendChild to handle the case of a -1 index to InsertChildNode, removes the InsertChildNode return value which is unused and cleans up various other things. Change-Id: I3a022e4dc2afffa6893ad11014034dd7ed301f13 Reviewed-on: https://pdfium-review.googlesource.com/26510 Commit-Queue: dsinclair <dsinclair@chromium.org> Reviewed-by: Tom Sepez <tsepez@chromium.org>
2018-02-13Cleanup CXFA_Node ownershipDan Sinclair
Currently a CXFA_Node can be owned by the CXFA_Document root pointer, the CXFA_Document Purge list or as a child of another CXFA_Node. This makes it hard to know who currently owns what. This CL moves all ownership of nodes into a CXFA_NodeOwner class which CXFA_Document subclasses. The node owner always owns all nodes and is responsible for cleaning them up upon destruction. The destruction order is not guarenteed to be in tree order as nodes can be inserted and moved around the tree. Change-Id: I6b202b8e844999ba835093dcdd1a808938b6d9a8 Reviewed-on: https://pdfium-review.googlesource.com/26434 Reviewed-by: Tom Sepez <tsepez@chromium.org> Commit-Queue: dsinclair <dsinclair@chromium.org>
2018-02-13Use actual type in place of void* in V8 mapTom Sepez
Add an override so we can comment about the need for it someday. Change-Id: I344c6fd7a635aeab030f0e769bd03a86406ffdd8 Reviewed-on: https://pdfium-review.googlesource.com/26310 Reviewed-by: dsinclair <dsinclair@chromium.org> Commit-Queue: Tom Sepez <tsepez@chromium.org>
2018-02-12Test if GC'd FXJS objects have their C-side counterparts cleaned up.Tom Sepez
Small correctness fix in other engine embeddertest. Change-Id: I6c2721921a659eef1b2f155ea1797722d37209d0 Reviewed-on: https://pdfium-review.googlesource.com/26270 Reviewed-by: dsinclair <dsinclair@chromium.org> Commit-Queue: Tom Sepez <tsepez@chromium.org>
2018-02-10Renable -Wimplicit-fallthrough for PDFium.Nico Weber
Didn't find any bugs, but also doesn't require much boilerplate. Bug: TODO Change-Id: I5f30183c55216f1333c79fcb46476d281720873e Reviewed-on: https://pdfium-review.googlesource.com/26210 Reviewed-by: Lei Zhang <thestig@chromium.org> Commit-Queue: Nico Weber <thakis@chromium.org>
2018-02-09Tidy V8TemplateMap.Tom Sepez
Change-Id: Idcba60bb36c0d47c0d2acb21ddfa3b3b46cdfe49 Reviewed-on: https://pdfium-review.googlesource.com/26170 Commit-Queue: Tom Sepez <tsepez@chromium.org> Reviewed-by: dsinclair <dsinclair@chromium.org>
2018-02-09Rename fxjs_v8.{h,cpp} to cfxjs_engine.{h,cpp}Tom Sepez
Place the template map definitions entirely in .cpp file. Change-Id: I2643f1b99f5582b69aa985857c4aa6f9b5ab57c8 Reviewed-on: https://pdfium-review.googlesource.com/26150 Commit-Queue: Tom Sepez <tsepez@chromium.org> Reviewed-by: dsinclair <dsinclair@chromium.org>
2018-02-09Move CFX_V8 testing from embeddertest to unittest.Tom Sepez
Test the lowest layer without firing up the whole library. Requires firing up v8 in the unit test main, though. Move array buffer allocator to cfx_v8 to allow building isolates. Change-Id: I9a56d503a48e0e555d3310f2997fa12137695860 Reviewed-on: https://pdfium-review.googlesource.com/26130 Commit-Queue: Tom Sepez <tsepez@chromium.org> Reviewed-by: dsinclair <dsinclair@chromium.org>
2018-02-08Move cjs_v8.cpp to cfx_v8.cppTom Sepez
The layering should be CJS => CFXJS => CFX_V8 with the CJS name being higher up. Change-Id: Ic130f248906e9c4df641dd508389b0555786b999 Reviewed-on: https://pdfium-review.googlesource.com/26051 Commit-Queue: Tom Sepez <tsepez@chromium.org> Reviewed-by: dsinclair <dsinclair@chromium.org>
2018-02-08Move XFA-specific methods out of CJS_V8.Tom Sepez
These belong more properly in CFXJS_Engine, and a small amount of casting is required to get to the sub-class. Change-Id: Id954d182b7a7d51fe0a522c04a50a1b362746c72 Reviewed-on: https://pdfium-review.googlesource.com/26050 Reviewed-by: dsinclair <dsinclair@chromium.org> Commit-Queue: Tom Sepez <tsepez@chromium.org>
2018-02-08Make CFXJS_Engine::SetObjectPrivate() static.Tom Sepez
Avoids call to CFXJS_Engine::EngineFromIsolateCurrentContext() during the Dispose() path, which feels scary because there aren't guarantees about it having an engine at isolate "dispose" time. Fortunately, |this| is not used, so make that fact clear. Replace some c-style callbacks with std::function while we're at it. Change-Id: Ia1a1a1fcc085d8657939e6f8c8d34fc511afddfe Reviewed-on: https://pdfium-review.googlesource.com/25970 Commit-Queue: Tom Sepez <tsepez@chromium.org> Reviewed-by: dsinclair <dsinclair@chromium.org>
2018-02-07Split creation of ordinary object and bound objects in FXJS.chromium/3343Tom Sepez
One can be performed by the CJS_V8 layer, the other requires the full FXJS mechanism. Avoids using -1 as a special case. Change-Id: I4a14ccb6a7fea393f84b70a07ada03b1a83c7d36 Reviewed-on: https://pdfium-review.googlesource.com/25830 Reviewed-by: dsinclair <dsinclair@chromium.org> Commit-Queue: Tom Sepez <tsepez@chromium.org>
2018-02-06Avoid needless malloc for v8:Global array.Tom Sepez
They have move semantics, so they can directly be members of the array. Change-Id: Ia73cbc06fd22665693ae1a4efe6a6f64eca1b5fb Reviewed-on: https://pdfium-review.googlesource.com/25490 Reviewed-by: dsinclair <dsinclair@chromium.org> Commit-Queue: Tom Sepez <tsepez@chromium.org>
2018-02-06Remove unused FreeObjectPrivate() overload.Tom Sepez
Change-Id: Idb1f6879a78e4ada8129250eedd37142b6f26588 Reviewed-on: https://pdfium-review.googlesource.com/25530 Reviewed-by: dsinclair <dsinclair@chromium.org> Commit-Queue: Tom Sepez <tsepez@chromium.org>
2018-02-05Fold CJS_EmbedObj classes into CJS_Object classesDan Sinclair
This CL removes the CJS_EmbedObj class and various subclasses and folds the subclasses into their CJS_Object counterparts. Change-Id: If6b882a4995c0b1bf83ac783f5c27ba9216c2d5c Reviewed-on: https://pdfium-review.googlesource.com/25410 Commit-Queue: dsinclair <dsinclair@chromium.org> Reviewed-by: Tom Sepez <tsepez@chromium.org>
2018-02-05Remove the CJS_EmbedObj template param from JSConstructor.Dan Sinclair
Each of the CJS_Objects can create their CJS_EmbedObj's internally and we don't need to do it though the JSConstructor. This also removes the need for the SetEmbedObj method in CJS_Object. Change-Id: Ib0535ad922b370634fd1e622a04860a96c4f2825 Reviewed-on: https://pdfium-review.googlesource.com/25370 Reviewed-by: Tom Sepez <tsepez@chromium.org> Commit-Queue: dsinclair <dsinclair@chromium.org>
2018-02-05Use unique pointer in CFXJS_PerObjectData.Tom Sepez
Also use the actual type information, not void* and remove casts. Template function not required to wrap virtual dtors. Change-Id: I9397cae136c3c395a368a1ef0ce8162d9b586076 Reviewed-on: https://pdfium-review.googlesource.com/25290 Reviewed-by: dsinclair <dsinclair@chromium.org> Commit-Queue: Tom Sepez <tsepez@chromium.org>
2018-02-05Fix some formcalc constant naming.Lei Zhang
Mark one constant as actually constant. Change-Id: Icf046dbd6751d07d6652ce24e4217a717a58b699 Reviewed-on: https://pdfium-review.googlesource.com/25310 Commit-Queue: Lei Zhang <thestig@chromium.org> Reviewed-by: dsinclair <dsinclair@chromium.org>
2018-02-05[XFA] dot_accessor may not provide a valid object.Dan Sinclair
If the object being accessed through the dot_accessor does not exist it's possible to receive a nullptr in ResolveObject. Instead of ASSERTing just return false and let the caller handle the failure. Bug: chromium::778904 Change-Id: I918c1c5dcb23b86082793c5037d58e0ab043559e Reviewed-on: https://pdfium-review.googlesource.com/25350 Reviewed-by: Henrique Nakashima <hnakashima@chromium.org> Reviewed-by: Ryan Harrison <rharrison@chromium.org> Commit-Queue: Ryan Harrison <rharrison@chromium.org>
2018-02-05Move const arrays to CFXJS_Engine from CJS_V8Tom Sepez
CJS_V8 is shared by both the non-xfa and xfa-side JS engines, but the const arrays are only used by the non-xfa side. Change-Id: Ic5ed8238df1a14dde8a4463b2388b4a7a923e392 Reviewed-on: https://pdfium-review.googlesource.com/25250 Commit-Queue: dsinclair <dsinclair@chromium.org> Reviewed-by: dsinclair <dsinclair@chromium.org>
2018-02-02Small tidy of fxjsTom Sepez
Use correct include style for v8 headers. Use incomplete type to avoid including header. Change-Id: Ibd58aa25b2a17fbd1cf6a20643b6f25ca008974f Reviewed-on: https://pdfium-review.googlesource.com/25230 Reviewed-by: Lei Zhang <thestig@chromium.org> Commit-Queue: Tom Sepez <tsepez@chromium.org>
2018-02-02Remove V8 context from CJS_V8Tom Sepez
Makes CFXJS_Engine and CFXJSE_Engine consistent with each other in that both have a V8 context to themselves, and not one inheritted from the CJS_V8 (which is now more per-isolate than per-context). Consolidate NewLocalContext() and GetPersistentContext(), which both did the exact same thing under the covers once inside v8 land. Rename a few things to make it simpler. Change-Id: I68905db9ad44253063da235fcb276a75627a2dbc Reviewed-on: https://pdfium-review.googlesource.com/25170 Reviewed-by: dsinclair <dsinclair@chromium.org> Commit-Queue: Tom Sepez <tsepez@chromium.org>
2018-02-02Make global object function names clearer.Tom Sepez
Also tidy some sub-expressions. Change-Id: Ieabd5f6cea60e8ec03c8ce5ebe372fc80b05a7bb Reviewed-on: https://pdfium-review.googlesource.com/25150 Reviewed-by: dsinclair <dsinclair@chromium.org> Commit-Queue: Tom Sepez <tsepez@chromium.org>
2018-02-01Make CFXJSE-created contexts refer to their CFXJS counteparts.chromium/3337Tom Sepez
Bug: 773229 Change-Id: Ic3774c7f6abe3a195bbe09b91d91c549d4d7ac46 Reviewed-on: https://pdfium-review.googlesource.com/25110 Reviewed-by: dsinclair <dsinclair@chromium.org> Commit-Queue: Tom Sepez <tsepez@chromium.org>
2018-02-01Change internal usages of GetPageIndexDeprecated() to new method.Henrique Nakashima
The new GetDestPageIndex() method does the same thing, but has a consistent form of returning an error (returns -1). Bug: pdfium:938 Change-Id: I31583e1c544d9173a28582b849edd5f73c40e174 Reviewed-on: https://pdfium-review.googlesource.com/25070 Reviewed-by: Lei Zhang <thestig@chromium.org> Commit-Queue: Henrique Nakashima <hnakashima@chromium.org>
2018-02-01Introduce CXFJS_Engine::EngineFromContext/SetEngineInContext helpersTom Sepez
No functional change, but expose helpers likely to be needed by next steps. Rename CXFJS_Engine::CurrentEngineFromIsolate to be more accurately named along the way. Change-Id: Iaf2d15a09cfed85705d24ee056e78edf0660ce59 Reviewed-on: https://pdfium-review.googlesource.com/25090 Reviewed-by: dsinclair <dsinclair@chromium.org> Commit-Queue: Tom Sepez <tsepez@chromium.org>
2018-02-01Make FXJSE_Engine constructed from FXJS_Engine.Tom Sepez
No change in functionality yet, just passing higher level object. Precursor to maybe sharing v8 context between fxjs / fxjse. Mark unimplemented ctors / assignment operator as "delete". Change-Id: I100de7755909eec2eed96f6f51216d85923ffbb2 Reviewed-on: https://pdfium-review.googlesource.com/25050 Commit-Queue: Tom Sepez <tsepez@chromium.org> Reviewed-by: dsinclair <dsinclair@chromium.org>
2018-02-01Deprecate FPDFDest_GetPageIndex and create a fixed version.Henrique Nakashima
FPDFDest_GetDestPageIndex() has a well defined return value for errors (-1). Keeping FPDFDest_GetPageIndex() to avoid changing behavior of the old API for whoever relies on it. Bug: pdfium:938 Change-Id: Iad528923cb156e957a419540c262a65f45cb777d Reviewed-on: https://pdfium-review.googlesource.com/24811 Commit-Queue: Henrique Nakashima <hnakashima@chromium.org> Reviewed-by: Lei Zhang <thestig@chromium.org>
2018-02-01Rename some CJS / IJS names for clarity.Tom Sepez
Prior cleanup before making FXJS/FXJSE share one v8 context. Return CJS object rather than isolate in one place. Use unique_ptr in one place. Change-Id: I837ae4880368a6d72e59b38f37e06908e05c34bd Reviewed-on: https://pdfium-review.googlesource.com/24950 Commit-Queue: dsinclair <dsinclair@chromium.org> Reviewed-by: dsinclair <dsinclair@chromium.org>
2018-02-01Test more context overlap in fxjs/fxjs_v8_embeddertest.cppTom Sepez
Change-Id: If9ab65019f12244938116e19f717e2092ccae4ff Reviewed-on: https://pdfium-review.googlesource.com/24931 Commit-Queue: dsinclair <dsinclair@chromium.org> Reviewed-by: dsinclair <dsinclair@chromium.org>
2018-01-31Remove handrolled search from GetEventParaInfoByNameRyan Harrison
BUG=pdfium:798 Change-Id: If6219f06303c78ecc8ddda52dd0095d658148e91 Reviewed-on: https://pdfium-review.googlesource.com/24718 Commit-Queue: Ryan Harrison <rharrison@chromium.org> Reviewed-by: dsinclair <dsinclair@chromium.org>
2018-01-31Remove NumericLimit params which are always the sameDan Sinclair
This CL removes the two params from NumericLmits which are always set to the same value and retrieves the values in the method itself. Change-Id: Ibeabdc7187faa58e93ffbe00063171fec0842256 Reviewed-on: https://pdfium-review.googlesource.com/24717 Reviewed-by: Henrique Nakashima <hnakashima@chromium.org> Reviewed-by: Ryan Harrison <rharrison@chromium.org> Commit-Queue: dsinclair <dsinclair@chromium.org>
2018-01-30Remove bare new from JS_Define.hTom Sepez
Change-Id: Ic734c7141e52b8ca332ca9dc78118b84904c41b0 Reviewed-on: https://pdfium-review.googlesource.com/24470 Reviewed-by: dsinclair <dsinclair@chromium.org> Commit-Queue: Tom Sepez <tsepez@chromium.org>
2018-01-30Cleanup some param passing codeDan Sinclair
This CL cleans up some of the code around passing input information around. Change-Id: Id3c1ffa93c9e50c3b6312b15533ccc32c7406264 Reviewed-on: https://pdfium-review.googlesource.com/24710 Reviewed-by: Tom Sepez <tsepez@chromium.org> Commit-Queue: dsinclair <dsinclair@chromium.org>
2018-01-29Fold single caller methods into callsiteDan Sinclair
This CL folds the GetNodeSameClassIndex, GetNodeSameNameIndex from CXFA_Node into the CJX_Tree callsites. Change-Id: Ia4f7750543e786ddd82abb4b6b4bde1bf1710f19 Reviewed-on: https://pdfium-review.googlesource.com/24471 Reviewed-by: Ryan Harrison <rharrison@chromium.org> Commit-Queue: dsinclair <dsinclair@chromium.org>
2018-01-29Kill some bare |new|s in cjs_field.cppTom Sepez
Change-Id: Id4d9f21d94d8b21199846ec5f7613590fd20579e Reviewed-on: https://pdfium-review.googlesource.com/24310 Reviewed-by: Lei Zhang <thestig@chromium.org> Commit-Queue: Tom Sepez <tsepez@chromium.org>
2018-01-29Cleanup the FFWidget determination code.Dan Sinclair
This CL converts the UIType to an XFA_FFWidgetType instead of reusing the XFA_Element type. The creation code is cleaned up to make it clearer what's happening. Change-Id: I5d3e4967d5c8b8a50dbb25e574b0d31fe0cf407a Reviewed-on: https://pdfium-review.googlesource.com/24390 Reviewed-by: Ryan Harrison <rharrison@chromium.org> Commit-Queue: dsinclair <dsinclair@chromium.org>
2018-01-29Fix identical if/else blocks.Lei Zhang
The CPWL_EditImpl::Backspace() instance occurred in commit 671a799. The other two instances where known cases. BUG=chromium:805881 Change-Id: Iae7cb8534e85342069d8753ced88543dad751e92 Reviewed-on: https://pdfium-review.googlesource.com/24050 Commit-Queue: dsinclair <dsinclair@chromium.org> Reviewed-by: dsinclair <dsinclair@chromium.org>
2018-01-24Cleanup CreateUIChild node creationDan Sinclair
This CL cleans up some of the code around creating the XML nodes related to the UI widgets. Change-Id: Ib91364439ab039f46e44690e92cc0cb93a8da203 Reviewed-on: https://pdfium-review.googlesource.com/23770 Reviewed-by: Henrique Nakashima <hnakashima@chromium.org> Commit-Queue: dsinclair <dsinclair@chromium.org>
2018-01-24Fix broken XFA builds with new V8Ryan Harrison
These two methods need to be switched to using the maybe versions, since the older versions are being deprecated. BUG=pdfium:989 Change-Id: Id2b48395353af0078e51b2d3ac42103643922320 Reviewed-on: https://pdfium-review.googlesource.com/23830 Reviewed-by: Ryan Harrison <rharrison@chromium.org> Commit-Queue: Ryan Harrison <rharrison@chromium.org>
2018-01-23Remove GetWidgetAcc callsDan Sinclair
This CL removes the calls to CXFA_Node::GetWidgetAcc() as they redundantly return the node they were called upon. Change-Id: I46e66cf98137a1dee7cd3fa8bc7d379eb97fded5 Reviewed-on: https://pdfium-review.googlesource.com/23630 Reviewed-by: Henrique Nakashima <hnakashima@chromium.org> Commit-Queue: dsinclair <dsinclair@chromium.org>
2018-01-23Merge CXFA_WidgetAcc into CXFA_Nodechromium/3330Dan Sinclair
This CL merges CXFA_WidgetAcc into CXFA_Node. Change-Id: Iab52ca7f98c9dcd4ab089c61b8f1324a565bf647 Reviewed-on: https://pdfium-review.googlesource.com/23610 Reviewed-by: Henrique Nakashima <hnakashima@chromium.org> Commit-Queue: dsinclair <dsinclair@chromium.org>
2018-01-23Use accessor for WidgetAccDan Sinclair
This CL removes places where we store CXFA_WidgetAcc into a variable and uses the accessor from CXFA_Node instead. This will ease the removal of CXFA_WidgetAcc. Change-Id: I7a6cb0079bb119481a5ae60679a537799b3be2c9 Reviewed-on: https://pdfium-review.googlesource.com/23510 Reviewed-by: Ryan Harrison <rharrison@chromium.org> Commit-Queue: dsinclair <dsinclair@chromium.org>
2018-01-23Tidy fxjs/cfxjse_isolatetracker.hTom Sepez
Move constructors to .cpp file. Remove unused member and its accessor method. Remove unused system includes. IsolateHandleRootContext is-a IsolateHandle (not has-a). Change-Id: Ifcba44f3ea7ad963fee731c24b2e79c5a8124539 Reviewed-on: https://pdfium-review.googlesource.com/23171 Commit-Queue: Tom Sepez <tsepez@chromium.org> Reviewed-by: Lei Zhang <thestig@chromium.org>