summaryrefslogtreecommitdiff
path: root/xfa/fxfa/fm2js/cxfa_fmparser_unittest.cpp
AgeCommit message (Collapse)Author
2018-10-04Reject argument lists that are not comma separatedRyan Harrison
The FormCalc grammar explicitly calls out that argument lists must have commas separating the simple expressions that make up the elements. The current implementation will accept the invalid string !a!b!c, which is 3 variables; !a, !b, and !c. BUG=chromium:890407 Change-Id: I3e2da4abce9989e9e9b929ce2da030e0f8dfd371 Reviewed-on: https://pdfium-review.googlesource.com/c/43430 Reviewed-by: Henrique Nakashima <hnakashima@chromium.org> Commit-Queue: Ryan Harrison <rharrison@chromium.org>
2018-06-12[fm2js] Convert formcalc tests to use raw literalsDan Sinclair
This CL updates some of the formcal tests to use raw literals instead of escaping lots of things. Change-Id: I539063b6c98aae318147b7103a0fd84e6b76054f Reviewed-on: https://pdfium-review.googlesource.com/34190 Commit-Queue: dsinclair <dsinclair@chromium.org> Reviewed-by: Tom Sepez <tsepez@chromium.org>
2018-06-06[fm2js] Fixup initialized variable conversionDan Sinclair
When converting a Var statement from formcalc to JS we would neglect to add the ; if the variable was initialized. This generated invalid JS as we'd end up with two statements smushed together. FormCalc: var s = "" Previous JS: var s = ""s = pfm.... New JS: var s = ""; s = pfm ... Bug: pdfium:1097 Change-Id: I8a869f07374cac68a06a487dace89699a4e5540c Reviewed-on: https://pdfium-review.googlesource.com/34110 Commit-Queue: dsinclair <dsinclair@chromium.org> Reviewed-by: Ryan Harrison <rharrison@chromium.org> Reviewed-by: Henrique Nakashima <hnakashima@chromium.org>
2018-05-16Convert Formcalc ToJavascript to a pointer WideTextBufDan Sinclair
This CL converts the WidetextBuf reference to a pointer in the formcalc ToJavaScript methods. Change-Id: Ia05c7255a99a2eaa3d9a57f77580969896ad90a0 Reviewed-on: https://pdfium-review.googlesource.com/32612 Reviewed-by: Henrique Nakashima <hnakashima@chromium.org> Commit-Queue: dsinclair <dsinclair@chromium.org>
2018-05-10Cleanup formcalc method generationchromium/3427Dan Sinclair
This CL simplifies the FormCalc method call generation when converted to JavaScript. Currently we output the same chunk of code to run the given method on an array or object per method call. This CL pulls out the common execution code to a pfm_method_runner function which is used instead. An embedder test has been added to verify that method invocation from formcalc works correctly. Bug: chromium:814848 Change-Id: I1ec052eab051053fedcb464d57e0e15228b8c5a2 Reviewed-on: https://pdfium-review.googlesource.com/32372 Reviewed-by: Henrique Nakashima <hnakashima@chromium.org> Commit-Queue: dsinclair <dsinclair@chromium.org>
2018-05-10Change formcalc javascript method outputDan Sinclair
This CL changes the formcalc JS conversion code so that method calls no longer generate exponential amounts of JavaScript. Previously we'd duplicate the code to call into a method twice. This would then generate twice once for arrays and once for other types of object. This CL changes the code to wrap the actual method call into a JavaScript function which will be used from both the array and non-array calling code. For the referenced bug, the generated JS originally needed a buffer of 365meg to generate. With this CL, it needs a buffer of 7.5k. Bug: chromium:814840 Change-Id: Ibb5993fa52b7c13b20b325cf8848a306f82ae014 Reviewed-on: https://pdfium-review.googlesource.com/32312 Reviewed-by: Henrique Nakashima <hnakashima@chromium.org> Reviewed-by: Ryan Harrison <rharrison@chromium.org> Commit-Queue: dsinclair <dsinclair@chromium.org>
2018-03-07[formcalc] Consider width along with depth of treeDan Sinclair
When building the formcalc parser trees we need to limit on width along with depth. It's possible to generate a tree of a single depth but is more then 20k nodes wide. This will eventuall cause stack overflow issues. This CL re-uses the depth check for the grammar expressions in which we're extending the width of the tree we count that against our depth check. Bug: chromium:813346 Change-Id: I01f6567a75776a75374465eacc1ff546db46cac1 Reviewed-on: https://pdfium-review.googlesource.com/28170 Reviewed-by: Ryan Harrison <rharrison@chromium.org> Reviewed-by: Henrique Nakashima <hnakashima@chromium.org> Commit-Queue: dsinclair <dsinclair@chromium.org>
2018-03-07[formcalc] Handle bad elseif conditionalschromium/3365Dan Sinclair
This Cl adds checking for the conditionals of if and elseif expressions. If the conditional fails to parse we should return nullptr. This already happens by accident in the if() case, but with elseif() conditions we'll fail the ASSERT in the CXFA_FMIfExpression constructor and crash. This CL explicitly checks for the expressions and early exists if they failed to parse. Bug: chromium:819509 Change-Id: I9a90182c7709c8c4c0d3ae17d6be67cb668c0c6a Reviewed-on: https://pdfium-review.googlesource.com/28131 Commit-Queue: dsinclair <dsinclair@chromium.org> Commit-Queue: Ryan Harrison <rharrison@chromium.org> Reviewed-by: Henrique Nakashima <hnakashima@chromium.org> Reviewed-by: Ryan Harrison <rharrison@chromium.org>
2018-03-05[formcalc] Simplify function signature parsingchromium/3363Dan Sinclair
This CL refactors the signature parsing for formcalc functions. Change-Id: I69634968ab1cbd9698a3ce3ae321e14bafe918c6 Reviewed-on: https://pdfium-review.googlesource.com/27930 Commit-Queue: dsinclair <dsinclair@chromium.org> Reviewed-by: Ryan Harrison <rharrison@chromium.org>
2018-03-05[formcalc] Cleanup m_error handlingDan Sinclair
This CL cleans up the setting of m_error. In most cases we don't need to set m_error it will be set when we bubble up the nullptr return from the various parse methods. The m_error was set inconsitently previously and was confusing on if it needed to be set or not. Change-Id: I8648b6296ef15239bd2663e6543a960b88177721 Reviewed-on: https://pdfium-review.googlesource.com/27910 Reviewed-by: Ryan Harrison <rharrison@chromium.org> Commit-Queue: dsinclair <dsinclair@chromium.org>
2018-02-22[formcalc] Cleanup function handlingDan Sinclair
This CL creates a CXFA_FMAST root for the AST tree instead of overloading the CXFA_FMFunctionDefinition. This Removes the m_global from FunctionDefinition and simpifies the code. Change-Id: I9347769a291ef1753539701f334cc8dd69b7187e Reviewed-on: https://pdfium-review.googlesource.com/27590 Commit-Queue: dsinclair <dsinclair@chromium.org> Reviewed-by: Ryan Harrison <rharrison@chromium.org>
2018-02-21[formcalc] Cleanup if expression parsing.Dan Sinclair
This CL simplifies the if expression parsing. The elseif construct has been split into its own list of IfExpressions and no longer gets stuffed into the elseExpression. A loop was added into the output methods for the elseif expressions, which also means we can go from having: if () { } else { if () { } else { } } to if () { } else if () { } else { } Which is a bit nicer. Change-Id: I6edf75215500d6a32a8d7218d477b6960f03de51 Reviewed-on: https://pdfium-review.googlesource.com/27571 Reviewed-by: Ryan Harrison <rharrison@chromium.org> Commit-Queue: dsinclair <dsinclair@chromium.org>
2018-02-21[formcalc] Add elseif expression to parser testDan Sinclair
This extends the test case for slightly more coverage. Change-Id: I97613c2d8df416b37718ff747d218dd61ea02b37 Reviewed-on: https://pdfium-review.googlesource.com/27570 Commit-Queue: dsinclair <dsinclair@chromium.org> Commit-Queue: Ryan Harrison <rharrison@chromium.org> Reviewed-by: Ryan Harrison <rharrison@chromium.org>
2018-02-20[formcalc] Treat elseif and else the same during outputDan Sinclair
This CL combines the elseif and else output code for formcalc. This causes an extra set of {}'s to be output in the else case but should be syntactically the same. Change-Id: Ie30a51efadae2034d117f4bf280b9f7abc950a23 Reviewed-on: https://pdfium-review.googlesource.com/27431 Commit-Queue: Ryan Harrison <rharrison@chromium.org> Reviewed-by: Lei Zhang <thestig@chromium.org> Reviewed-by: Ryan Harrison <rharrison@chromium.org>
2018-02-20Tighten up assignment instructions in formcalc.Dan Sinclair
The assignment operator can not be chained. This Cl removes the while loop for assignments and changes it to an if(). We also can not have an assignment inside ()'s so remove that option. Bug: chromium:779349 Change-Id: I6934e18815f843ae8241023df6c03d8bbcd8168d Reviewed-on: https://pdfium-review.googlesource.com/27350 Commit-Queue: Ryan Harrison <rharrison@chromium.org> Reviewed-by: Ryan Harrison <rharrison@chromium.org>
2018-02-19Merge FormCalc to JavaScript methodsdan sinclair
This CL merges the ToJavaScript and the ToImpliedJS methods in the FormCalc expressions. The type of return is passed as a paramter to ToJavaScript. Change-Id: Idff83677bc70b964d95aa6ff6b0e2c1bf8c603ea Reviewed-on: https://pdfium-review.googlesource.com/27210 Reviewed-by: Ryan Harrison <rharrison@chromium.org> Commit-Queue: dsinclair <dsinclair@chromium.org>
2017-10-25Add in depth check for ToJavascript and related methodsRyan Harrison
There exists a similar check for the parser, but it doesn't catch all cases of excessive memory usage, since a single parse step can generate multiple expressions that need to be converted or other cases where the parse depth doesn't match the emission depth later. Due to the expressions appearing in two different inheritence hierachies the depth information needs to be stored outside of the classes, thus the new depth class. Another way to handle this would be to change the method calls to take in a visitor object that tracks depth. This would require significant reworking of some of the code, so I am going to file a bug about doing that conversion as a cleanup. BUG=chromium:752495 Change-Id: Ica7c9b60ecf1e17530ea88b7bfb01582c63043be Reviewed-on: https://pdfium-review.googlesource.com/16752 Commit-Queue: Ryan Harrison <rharrison@chromium.org> Reviewed-by: Henrique Nakashima <hnakashima@chromium.org>
2017-09-18Convert string class namesRyan Harrison
Automated using git grep & sed. Replace StringC classes with StringView classes. Remove the CFX_ prefix and put string classes in fxcrt namespace. Change AsStringC() to AsStringView(). Rename tests from TEST(fxcrt, *String*Foo) to TEST(*String*, Foo). Couple of tests needed to have their names regularlized. BUG=pdfium:894 Change-Id: I7ca038685c8d803795f3ed02545124f7a224c83d Reviewed-on: https://pdfium-review.googlesource.com/14151 Reviewed-by: Tom Sepez <tsepez@chromium.org> Commit-Queue: Ryan Harrison <rharrison@chromium.org>
2017-08-30Move CFX_WideTextBuf out of fx_basicDan Sinclair
This CL moves CFX_WideTextBuf to its own files and updates includes as needed. Change-Id: Ibe66ecf3e66f8f01dd8e9eaf6b467588be86ad4f Reviewed-on: https://pdfium-review.googlesource.com/12413 Reviewed-by: Tom Sepez <tsepez@chromium.org> Commit-Queue: dsinclair <dsinclair@chromium.org>
2017-08-29Fix incorrectly guarded lexer advancement in FormCalc parserRyan Harrison
BUG=chromium:752501 Change-Id: Ie9943cd80a4afa73ad9393c8bcd2aa2656a9d932 Reviewed-on: https://pdfium-review.googlesource.com/12290 Commit-Queue: dsinclair <dsinclair@chromium.org> Reviewed-by: dsinclair <dsinclair@chromium.org>
2017-08-16Add parse depth limit to FormCalc parserRyan Harrison
Due to the recursive nature of the FormCalc parser, deeply nested expressions can lead to memory being exhausted. This check is being added to have the parser exit early instead of running out of memory. This should reduce the number of false positives about addressing issues being found by fuzzers. BUG=chromium:752433 Change-Id: I511ecfb07e32073555e1fd1658f3b8b47f1a5a91 Reviewed-on: https://pdfium-review.googlesource.com/11170 Commit-Queue: Ryan Harrison <rharrison@chromium.org> Reviewed-by: Tom Sepez <tsepez@chromium.org>
2017-07-27Move CXFA_FMParse to CFXA_FMParserRyan Harrison
Minor nit I had with this code and want to fix. Bug: Change-Id: I2494667d580bf982a4359153123d9e8afb75e349 Reviewed-on: https://pdfium-review.googlesource.com/9370 Reviewed-by: Tom Sepez <tsepez@chromium.org> Reviewed-by: (OOO Jul 28 - Aug 8) dsinclair <dsinclair@chromium.org> Commit-Queue: Ryan Harrison <rharrison@chromium.org>