[for reference] all work done which is not in original repo #11

…in the lexer rules using them.

…and regex minification (see the jison and jison-lex repos for the work done on the latter subject)

…s still happen to deliver a LARGER regex string in the end, we also check against the plain, unadulterated regex set expressions.

…and trailing whitepace off an 'arrowed' action code snippet: it's supposed to be [the latter part of] a single line of code anyway.

…end a semicolon at the end of an 'arrowed' action code snippet: first, it *may* already have a semicolon at the end, which would have us produce sub-optimal code if we want this to pass code linting tools and second, this kind of code 'cleaning' should be done together with all the other 'complex code manipulation stuff' down in jison itself. Corrected the unit tests to match the new output.

… into leading-underscores

# Conflicts: # .gitignore # bnf.l # bnf.y # ebnf-transform.js # ebnf.y # parser.js # tests/ebnf.js # transform-parser.js

- added tests (some fail) to check proper operation when feeding parser literal tokens containing quotes of any kind. - `yytext.length` --> `yyleng` - regenerated parsers

…xp` as a dependency (required since `ebnf-parser` supports Unicode via the regexes in the lexer) # Conflicts: # package.json # parser.js # transform-parser.js

- quote police

…d 161 (hadn't run the proper `make git-tag` + `make bump` commands!)

…ions as name+value pairs. Adjusted the relevant unit tests accordingly.

…he b0rkb0rkb0rk once again. :-(

… construction was flaked as it still carried comma's from its antique `console.warn` heritage.

…ction to a rule. (`rule -> action`)

…x.html is not working yet, but that wasn't available before either, so nothing is lost. All tests pass in node via `make` or `make test`)

… they're indented and no empty lines in between. Also allow option values to be boolean `false`, `true` or any *finite* numeric value, e.g. `%options debug=3`

…equired (`%include` command)

… is now in sync with the main BNF parser.

…res: macros expand to regex closures, which are available via the `this.matches[]` member array in the lexer action code.

…sn't fetched via `this.matches[]` anyway: this should be slightly faster at run-time

…Also use the `%epsilon` in the grammars for better readability of empty rules.

…cky fix for an `npm install` issue on Windows

…code identifier, while the lexer+parser of the jison language already do: this meant any Unicode aliases in a EBNF grammar would break JISON.

…g. `expr expr` would turn into `expr[expr1] expr[expr2]` implicitly, but `expr3 expr` MUST NOT become aliased as `expr3` turns off all `expr`-basename aliases to prevent confusion. This same bug exists in JISON itself and is being fixed there as well. The way to observe this bug and make it cause undesired behaviour is using a rule where the implicit alias numbers would be 11 or greater, e.g. a rule like `a1 a1 a1 a1 a1 a1 a1 a1 a1 a1 a1 a11` which would otherwise be implicitly aliased as `a1[a11] a1[a12] a1[a13] a1[a14] a1[a15] a1[a16] a1[a17] a1[a18] a1[a19] a1[a110] a1[a111] a11[a111]`: note the undesirable clash between implicit alias `a11` which will block the direct term `a11` and thus produce incorrect action code when `$a11` is used in the action code. Now any term or alias with a number tail is properly marked up as blocking any creation of implicit aliases of the same base name. See also the comment chunk in the JISON.JS file, where this same logic originates as we have to duplicate that bit for the EBNF to BNF rewrite logic in this library: ``` var count = {}, names = {}, donotalias = {}; // When the rule is fitted with aliases it doesn't mean that the action code MUST use those: // we therefor allow access to both the original (non)terminal and the alias. // // Also note that each (non)terminal can also be uniquely addressed by [$@]<nonterminal><N> // where N is a number representing the number of this particular occurrence of the given // (non)terminal. // // For example, given this (intentionally contrived) production: // elem[alias] elem[another_alias] another_elem[alias] elem[alias] another_elem[another_alias] // all the items can be accessed as: // $1 $2 $3 $4 $5 // $elem1 $elem2 $another_elem1 $elem3 $another_elem2 // $elem $elem2 $another_elem $elem3 $another_elem2 // $alias1 $another_alias1 $alias2 $alias3 $another_alias2 // $alias $another_alias $alias2 $alias3 $another_alias2 // where each line above is equivalent to the top-most line. Note the numbers postfixed to // both (non)terminal identifiers and aliases alike and also note alias2 === another_elem1: // the postfix numbering is independent. // // WARNING: this feature is disabled for a term when there already exists an // (human-defined) *alias* for this term *or* when the numbered auto-alias already // exists because the user has used it as an alias for another term, e.g. // // e: WORD[e1] '=' e '+' e; // // would *not* produce the `e1` and `e2` aliases, as `e1` is already defined // as an explicit alias: adding auto-alias `e1` would then break the system, // while `e2` would be ambiguous from the human perspective as he *might* then // expect `e2` and `e3`. var addName = function addName(s) { var base = s.replace(/[0-9]+$/, ''); var dna = donotalias[base]; if (names[s]) { count[s]++; if (!dna) { names[s + count[s]] = i + 1; } } else { names[s] = i + 1; count[s] = 1; if (!dna) { names[s + count[s]] = i + 1; } } }; // register the alias/rule name when the real one ends with a number, e.g. `rule5` as // *blocking* the auto-aliasing process for the term of the same base, e.g. `rule`. // This will catch the `WORD[e1]` example above too, via `e1` --> `donotalias['e']` var markBasename = function markBasename(s) { if (/[0-9]$/.test(s)) { s = s.replace(/[0-9]+$/, ''); donotalias[s] = true; } }; for (i = 0; i < rhs.length; i++) { // mark both regular and aliased names, e.g., `id[alias1]` and `id1` rhs_i = aliased[i]; markBasename(rhs_i); if (rhs_i !== rhs[i]) { markBasename(rhs[i]); } } for (i = 0; i < rhs.length; i++) { // check for aliased names, e.g., id[alias] rhs_i = aliased[i]; addName(rhs_i); if (rhs_i !== rhs[i]) { addName(rhs[i]); } } ```

…, e.g. a `$name` named variable reference in your action code where `name` exists in multiple positions in your production rule, hence it would be ambiguous which one to pick (previously, the code picked the first occurrence in the production, but then there was the possibility of clashes of an identical alias and term name, where the aliased term would come first: then the `name` term position would be picked instead: confusion abounds, hence we now check and FAIL on ambiguous term references!

…mmon coding mistakes.

…ng from the lexer.

…ackage-lock.json file in the commit set :-(

… `nvm` labels in there any more (`stable` --> `node`) + update NPM packages

…e lexer, so this hacky code is not ever necessary anymore

…of console.log, console.warn, et al: https://nodejs.org/api/console.html#console_console_log_data_args - provide the same detailed error report for unknown/unsupported parser options as we do for unsupported lexer options in the lex-parser module!

… we want to use the bleeding edge jison code which should have installed this module as a git submodule. (TravisCI should merely use the jison-gho as listed in the package.json file though!)

… debug=1 in the test code's parser_reset() function: RTFC) - fix/clean a few tests.

…t source(s) and update the `build` make target accordingly.

…duced by new NPM. - bumped build revision. - rebuilt library and reference files.

…the BNF lexer and parser will have taken care of translating those to their standard values `$end` and `` (empty for %epsilon)

…eof` so that we can properly recognize them and translate these to the standard value `$end` as soon as possible.

…default action processing in jison) and regenerate library files

…old in jison: a few rules' actions didn't assign a value to `$$` as that was done *implicitly* by the default action code preceding every action rule up to now. Since jison commit GerHobbelt/jison@5f8e197 this prelude DOES NOT EXIST any more and this type of nasty use of that old knowledge will FAIL from now on!

…ules' literal terms surrounded by double quotes by default (`dquote()` function); no changed functionality.

# Conflicts: # package.json

…ison tool itself, as jison and this and the other 'modules' are very tightly related in reality.

…*including cr/lf newlines**

…-` dashes and thus are not merely `ID`s but `NAME`s are also accepted here!

…e run-time for error reporting, etc.)

…ed error reporting later on...

… `%{...%}` almost everywhere internally, apply that to the new %code section as well.

…pmjs.com/misc/scope; this is also related to GerHobbelt/jison#11 -- rant: this need to change every bloody `require()` call in the code for a 'scoped package' is what I particularly dislike about this approach, but so far, it's the best we've got as long as the entire world hasn't upgraded to packge-lock.json support... /rant

… as per https://docs.npmjs.com/misc/scope; this is also related to GerHobbelt/jison#11 -- rant: this need to change every bloody `require()` call in the code for a 'scoped package' is what I particularly dislike about this approach, but so far, it's the best we've got as long as the entire world hasn't upgraded to packge-lock.json support... /rant

…ment (and using that one instead of the npm dependency!) in the Makefile

… the board

…ll development environment (and using that one instead of the npm dependency!) in the Makefile

…fail

…n and content for easier error inspection by the user. Uses the same code as the lex grammar parser in the jison lex-parser module.

… these can be arbitrary, non-whitespace values (unless formatted as true strings, surrounded within quotes) - just like `%debug`, let the *grammar* handle the `%ebnf` setting (as much as possible, anyway) - fix the issue where the *global variable* `ebnf` was created to track the `%ebnf` setting: parsers should not inject variables into the global scope, but use the `yy` shared store instead!

…ide with another (user-written) rule's symbol in the given grammar: added code which makes sure the generated symbol is unique, i.e. does not collide with any other rule symbols. **WARNING**: IFF you previously depended on **the undocumented feature how EBNF rule particles were expanded into BNF** for `$name`-referencing those parts of the rule's production in your action code, than those code chunks will be broken as from now on there is no guarantee that the first `*` expansion will be named `<symbol>_repetition1`, etc.! You SHOULD have used aliases instead, anyhow, e.g. `T*[Tstar] -> T_star` instead of `T* -> T_repetition1`.

…* `ebnf` was created to track the `%ebnf` setting: parsers should not inject variables into the global scope, but use the `yy` shared store instead!) ==> when the old code had this (admittedly nasty) parser-lexer communication hack (lexer is generated *inside* the parser so the `var ebnf` declared in bnf.y would be seen as 'in scope' closure variable for the generated lexer!) and you (that's **me**!) remove that half of the hack, make sure the other side picks up the 'ebnf' setting in time to drive `extend()` the way it was intended, once the grammar has been parsed!

… several temporary print debug statements the truth was uncovered: a simple omission.

…pass.

…and `##n`: Note that `#name` are straight **static** symbol translations, which are okay as they don't require access to the parse stack: `#n` references can be resolved completely at grammar compile time.

- rebuilt library files

… this cuts down significantly on duplicated code right now.

…atures that originate in those. Pushing an intermediate release now to ensure the build process will fly on the next one. GRMBL.

…printing, AST exports, etc.

…t as a constant and return the transformed grammar in a new object, which is otherwise a deep clone of the input.

…ammar AST.

…ng Zachary Carter, just like the package.json file does)

…or mocha to perform leak detection.

…%import statements!) - fix one error handling chunk which didn't properly reference the yylexer.prettyPrint API

…or mocha to perform leak detection.

…! (N.B.: this is why it was, ahhh, very unwise to split the jison project into multiple repo's, by the way. An approach like that done with babel (one repo, multiple NPM packages) would have made much more sense if you really want to compartmentalize jison into 'modules'...)

… used (and duplicated) all over the place.

…emplates and the `rmCommonWS()` utility API.

…ides *all* APIs.

…used until jison properly supports `%code imports %{...%}`

…duce ES6 and ES5 bundles in the `dist/` directory for external perusal.

…on-helpers-lib API

…ode block as coding validation.

…JISON again.

…he ES6 code generator migration

…lopment utility scripts

… undesirable as this is the secondary source repo!)

…e 'external dependencies' ARE NOT included in the rollup but kept external. ==> jison-helpers-lib + lex-parser MUST NOT be included in the dist/ rollup library files for ebnf-parser!

…ADME's

Commits on Nov 9, 2016

regenerated library

GerHobbelt committed Nov 9, 2016

Configuration menu

View commit details

Copy full SHA for bbdc68f

Browse repository at this point

Copy the full SHA

bbdc68f View commit details

Browse the repository at this point in the history

Commits on Nov 19, 2016

rebuild library files

GerHobbelt committed Nov 19, 2016

Configuration menu

View commit details

Copy full SHA for 62f1b4c

Browse repository at this point

Copy the full SHA

62f1b4c View commit details

Browse the repository at this point in the history

Commits on Dec 2, 2016

Create parser.js

mamartel authored Dec 2, 2016

Configuration menu

View commit details

Copy full SHA for 26d51c8

Browse repository at this point

Copy the full SHA

26d51c8 View commit details

Browse the repository at this point in the history

Commits on Dec 3, 2016

rebuild library files

GerHobbelt committed Dec 3, 2016

Configuration menu

View commit details

Copy full SHA for 1520203

Browse repository at this point

Copy the full SHA

1520203 View commit details

Browse the repository at this point in the history

Commits on Jan 28, 2017

bumped build number

GerHobbelt committed Jan 28, 2017

Configuration menu

View commit details

Copy full SHA for 34c785b

Browse repository at this point

Copy the full SHA

34c785b View commit details

Browse the repository at this point in the history

Commits on Apr 15, 2017

bump version

GerHobbelt committed Apr 15, 2017

Configuration menu

View commit details

Copy full SHA for 4643c6f

Browse repository at this point

Copy the full SHA

4643c6f View commit details

Browse the repository at this point in the history

Commits on Aug 8, 2017

regenerate library files

GerHobbelt committed Aug 8, 2017

Configuration menu

View commit details

Copy full SHA for b115cb7

Browse repository at this point

Copy the full SHA

b115cb7 View commit details

Browse the repository at this point in the history

Commits on Aug 28, 2017

whitespace police raid

GerHobbelt committed Aug 28, 2017

Configuration menu

View commit details

Copy full SHA for 759a611

Browse repository at this point

Copy the full SHA

759a611 View commit details

Browse the repository at this point in the history

Commits on Sep 10, 2017

made all unit tests pass once again

GerHobbelt committed Sep 10, 2017

Configuration menu

View commit details

Copy full SHA for de04d67

Browse repository at this point

Copy the full SHA

de04d67 View commit details

Browse the repository at this point in the history

Commits on Sep 17, 2017

rebuilt library files

GerHobbelt committed Sep 17, 2017

Configuration menu

View commit details

Copy full SHA for e6b12e5

Browse repository at this point

Copy the full SHA

e6b12e5 View commit details

Browse the repository at this point in the history

Commits on Oct 16, 2017

sync

GerHobbelt committed Oct 16, 2017

Configuration menu

View commit details

Copy full SHA for bb1a73b

Browse repository at this point

Copy the full SHA

bb1a73b View commit details

Browse the repository at this point in the history

Commits on Oct 24, 2017

obsoleted. point at the jison monorepo.

GerHobbelt committed Oct 24, 2017

Configuration menu

View commit details

Copy full SHA for 0cf82aa

Browse repository at this point

Copy the full SHA

0cf82aa View commit details

Browse the repository at this point in the history

[for reference] all work done which is not in original repo #11

Are you sure you want to change the base?

[for reference] all work done which is not in original repo #11

Commits on Sep 15, 2016

Commits on Sep 18, 2016

Commits on Sep 19, 2016

Commits on Nov 9, 2016

Commits on Nov 10, 2016

Commits on Nov 11, 2016

Commits on Nov 14, 2016

Commits on Nov 17, 2016

Commits on Nov 18, 2016

Commits on Nov 19, 2016

Commits on Dec 2, 2016

Commits on Dec 3, 2016

Commits on Dec 13, 2016

Commits on Dec 15, 2016

Commits on Jan 25, 2017

Commits on Jan 27, 2017

Commits on Jan 28, 2017

Commits on Jan 30, 2017

Commits on Jan 31, 2017

Commits on Feb 1, 2017

Commits on Feb 3, 2017

Commits on Feb 10, 2017

Commits on Feb 19, 2017

Commits on Feb 20, 2017

Commits on Feb 21, 2017

Commits on Mar 4, 2017

Commits on Mar 8, 2017

Commits on Mar 26, 2017

Commits on Apr 8, 2017

Commits on Apr 9, 2017

Commits on Apr 15, 2017

Commits on Jun 22, 2017

Commits on Jul 22, 2017

Commits on Jul 23, 2017

Commits on Jul 31, 2017

Commits on Aug 8, 2017

Commits on Aug 19, 2017

Commits on Aug 20, 2017

Commits on Aug 21, 2017

Commits on Aug 26, 2017

Commits on Aug 27, 2017

Commits on Aug 28, 2017

Commits on Aug 31, 2017

Commits on Sep 1, 2017

Commits on Sep 5, 2017

Commits on Sep 6, 2017

Commits on Sep 10, 2017

Commits on Sep 11, 2017

Commits on Sep 12, 2017

Commits on Sep 17, 2017

Commits on Sep 24, 2017

Commits on Sep 29, 2017

Commits on Oct 3, 2017

Commits on Oct 12, 2017

Commits on Oct 13, 2017

Commits on Oct 15, 2017

Commits on Oct 16, 2017

Commits on Oct 23, 2017

Commits on Oct 24, 2017

Commits on Dec 13, 2017