aboutsummaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* Make: Use $< over $?HEADmainEuAndreh2025-05-041-1/+1
|
* Combine "CompiledGrammar" and "Report" payloadsEuAndreh2024-12-214-66/+26
| | | | | | | | Instead of one JSON file for each output, a single file with both outputs combined, each under its own key. The beauty of maps. Also instead of using a flag to determine where to put the data, just print to stdout and be done with it.
* chmod +x mkdeps.shEuAndreh2024-12-121-0/+0
|
* Makefile: Uncomment dependency on build filesEuAndreh2024-12-121-3/+3
|
* tests/unit/: Actually run existing testsEuAndreh2024-12-1212-12/+231
|
* Build and run existing test executables with no actual testsEuAndreh2024-12-1227-45/+247
|
* Move existing test files to "urubu" subdirectoryEuAndreh2024-12-1113-0/+7
|
* Do the same single file consolidation on testsEuAndreh2024-12-1128-4292/+4182
|
* Consolidate packages spread across multiple files into single oneEuAndreh2024-12-1161-10246/+10003
|
* rm LICENSEEuAndreh2024-12-111-21/+0
|
* rm .github/workflows/test.ymlEuAndreh2024-12-111-32/+0
|
* rm go.mod go.sumEuAndreh2024-12-112-20/+0
|
* Namespace packages with "urubu/"EuAndreh2024-12-1093-189/+189
|
* Start building test filesEuAndreh2024-12-1022-47/+109
|
* cmd/: Simplify flag usageEuAndreh2024-12-108-189/+34
|
* Build and test existing code as isEuAndreh2024-12-0232-68/+155
|
* Setup Makefile project structure with empty src/urubu.go sourceEuAndreh2024-11-1920-0/+458
|
* Add byte position to nodes of a syntax treeRyo Nihei2022-11-134-5/+22
|
* Fix typoRyo Nihei2022-11-133-12/+12
|
* Add byte position to tokensRyo Nihei2022-11-132-241/+270
|
* Set token positions for the EOF symbolRyo Nihei2022-11-122-3/+3
| | | | | | vartan intentionally had not set positional information for the EOF symbol in a token because the corresponding character does not exist in the source code. However, to be able to include the positional information in a syntax error message when a syntax error occurs due to the detection of the EOF symbol during parsing, this commit sets the position next to the last character in the source code as the position of the EOF symbol.
* Fix counting of token positionsRyo Nihei2022-11-121-37/+46
| | | | The name `unread` gives the impression of operating just the opposite of `read`. However, this commit renames `unread` function to `revert` because it was a function to revert the state to the last acceptance.
* Update CHANGELOGRyo Nihei2022-11-111-0/+32
|
* Import source code of lexer generatorRyo Nihei2022-11-1081-1168/+17101
| | | | From: https://github.com/nihei9/maleeni
* Split SymbolTable's APIs into reader/writerRyo Nihei2022-11-066-132/+171
|
* Move the skip table from lexer-related data to parser-related dataRyo Nihei2022-11-068-81/+69
|
* Remove anonymous symbol systemRyo Nihei2022-11-0513-277/+233
| | | | Remove unimportant features to tidy up the specification.
* Remove alias systemRyo Nihei2022-11-0516-240/+54
| | | | Remove unimportant features to tidy up the specification.
* Update READMERyo Nihei2022-08-081-0/+2
|
* Upgrade Go compiler to v1.19Ryo Nihei2022-08-076-32/+34
|
* Add testsRyo Nihei2022-08-061-0/+14
|
* Remove underscore syntax matching any symbolRyo Nihei2022-08-063-59/+8
| | | | | | | | | Underscore syntax: For instance, a tree `(expr (id 'a') (add '+') (_))` matches both source codes `a + b * c` and `a - b / c`. This feature is helpful because it allows you to emphasize the main points of the test by ignoring nodes of no interest. However, we will remove the feature for the time being to reconsider the grammar.
* Prohibit error node having childrenRyo Nihei2022-08-052-0/+85
|
* Fix parse error messages for tree parserRyo Nihei2022-06-141-36/+54
|
* Support testing token's texts in vartan-test commandRyo Nihei2022-06-1310-193/+926
|
* Prohibit using a pattern in an alternativeRyo Nihei2022-06-126-19/+102
| | | | | | When a syntax error occurs, the parser must provide a user with the names of expected tokens. However, if a pattern appears directly in an alternative, Vartan cannot assign an appropriate name to the pattern. Therefore, this commit prohibits alternatives from containing patterns.
* Support the underscore symbol matching any symbols in vartan-test commandRyo Nihei2022-06-114-3/+55
|
* Remove the kind field from a node corresponding to an anonymous terminal symbolRyo Nihei2022-06-118-43/+249
|
* Support testable tree output in vartan-parse commandRyo Nihei2022-05-294-7/+66
|
* Add vartan-test commandRyo Nihei2022-05-2912-1/+2535
|
* Rename spec package to spec/grammar packageRyo Nihei2022-05-2729-29/+29
|
* Allows a directory to be specified as the --output option for the ↵Ryo Nihei2022-05-227-76/+112
| | | | vartan-compile command
* Stop supporting SLR(1) and always use LALR(1)Ryo Nihei2022-05-2220-1182/+67
|
* Prohibit applying #left, #right, #assign, and #prec to an error symbolRyo Nihei2022-05-152-14/+103
| | | | | | | The shift of the error symbol is an operation forced by the driver. Therefore it is impossible to change this behavior by giving precedence to the error symbol. If we desire to change the precedence of a production rule with the error symbol, we can use #prec directive.
* Prohibit applying the expansion operator to anything other than identifiersRyo Nihei2022-05-153-2/+62
|
* Prohibit using escape sequences in string literalsRyo Nihei2022-05-155-86/+58
|
* Stop handling panic to print a stack traceRyo Nihei2022-05-156-135/+41
|
* Update CHANGELOGRyo Nihei2022-05-101-0/+15
|
* Follow golangci-lintRyo Nihei2022-05-101-3/+1
|
* Add spelling inconsistencies checkRyo Nihei2022-05-103-23/+238
|