calmjs.parse 1.2.3

Introduction

For any kind of build system that operates with JavaScript code in conjunction with a module system, the ability to understand what modules a given set of sources require or provide is paramount. As the Calmjs project provides a framework that produces and consume these module definitions, the the ability to have a comprehensive understanding of given JavaScript sources is a given. This goal was originally achieved using slimit, a JavaScript minifier library that also provided a comprehensive parser class that was built using Python Lex-Yacc (i.e. ply).

However, as of mid-2017, it was noted that slimit remained in a minimum state of maintenance for more than four years (its most recent release, 0.8.1, was made 2013-03-26), along with a number of serious outstanding issues have left unattended and unresolved for the duration of that time span. As the development of the Calmjs framework require those issues to be rectified as soon as possible, a decision to fork the parser portion of slimit was made. This was done in order to cater to the interests current to Calmjs project at that moment in time.

The fork was initial cut from another fork of slimit (specifically lelit/slimit), as it introduced and aggregated a number of bug fixes from various sources. To ensure a better quality control and assurance, a number of problematic changes introduced by that fork were removed. Also, new tests were created to bring coverage to full, and issues reported on the slimit tracker were noted and formalized into test cases where applicable. Finally, grammar rules were updated to ensure better conformance with the ECMA-262 (ES5) specification.

The goal of calmjs.parse is to provide a similar API that slimit had provided, except done in a much more extensible manner with more correctness checks in place. This however resulted in some operations that might take longer than what slimit had achieved, such as the pretty printing of output.

A CLI front-end that makes use of this package is provided through crimp.

Installation

The following command may be executed to source the latest stable version of calmjs.parse wheel from PyPI for installation into the current Python environment.

$ pip install calmjs.parse

As this package uses ply, it requires the generation of optimization modules for its lexer. The wheel distribution of calmjs.parse does not require this extra step as it contains these pre-generated modules for ply up to version 3.11 (the latest version available at the time of previous release), however the source tarball or if ply version that is installed lies outside of the supported versions, the following caveats will apply.

If a more recent release of ply becomes available and the environment upgrades to that version, those pre-generated modules may become incompatible, which may result in a decreased performance and/or errors. A corrective action can be achieved through a manual optimization step if a newer version of calmjs.parse is not available, or ply may be downgraded back to version 3.11 if possible.

Once the package is installed, the installation may be tested or be used directly.

$ pip install git+https://github.com/calmjs/calmjs.parse.git#egg=calmjs.parse

$ python -m calmjs.parse.parsers.optimize

$ python -m unittest calmjs.parse.tests.make_suite

Usage

As this is a parser library, no executable shell commands are provided. There is however a helper callable object provided at the top level for immediate access to the parsing feature. It may be used like so:

>>> from calmjs.parse import es5
>>> program_source = u'''
... // simple program
... var main = function(greet) {
...     var hello = "hello " + greet;
...     return hello;
... };
... console.log(main('world'));
... '''
>>> program = es5(program_source)
>>> # for a simple repr-like nested view of the ast
>>> program  # equivalent to repr(program)
<ES5Program @3:1 ?children=[
  <VarStatement @3:1 ?children=[
    <VarDecl @3:5 identifier=<Identifier ...>, initializer=<FuncExpr ...>>
  ]>,
  <ExprStatement @7:1 expr=<FunctionCall @7:1 args=<Arguments ...>,
    identifier=<DotAccessor ...>>>
]>
>>> # automatic reconstruction of ast into source, without having to
>>> # call something like `.to_ecma()`
>>> print(program)  # equivalent to str(program)
var main = function(greet) {
  var hello = "hello " + greet;
  return hello;
};
console.log(main('world'));

>>>

Please note the change in indentation as the default printer has its own indentation scheme. If comments are needed, the parser can be called using with_comments=True:

>>> program_wc = es5(program_source, with_comments=True)
>>> print(program_wc)
// simple program
var main = function(greet) {
  var hello = "hello " + greet;
  return hello;
};
console.log(main('world'));

>>>

Also note that there are limitations with the capturing of comments as documented in the Limitations section.

The parser classes are organized under the calmjs.parse.parsers module, with each language being under their own module. A corresponding lexer class with the same name is also provided under the calmjs.parse.lexers module. For the moment, only ES5 support is implemented.

>>> from calmjs.parse.unparsers.es5 import pretty_print
>>> print(pretty_print(program, indent_str='    '))
var main = function(greet) {
    var hello = "hello " + greet;
    return hello;
};
console.log(main('world'));

>>>

>>> from calmjs.parse.unparsers.es5 import minify_print
>>> print(minify_print(program))
var main=function(greet){var hello="hello "+greet;return hello;};...
>>> print(minify_print(program, obfuscate=True, obfuscate_globals=True))
var a=function(b){var a="hello "+b;return a;};console.log(a('world'));

>>> # pretty print without comments being parsed
>>> print(es5.pretty_print(program_source))
var main = function(greet) {
  var hello = "hello " + greet;
  return hello;
};
console.log(main('world'));

>>> # pretty print with comments parsed
>>> print(es5.pretty_print(program_source, with_comments=True))
// simple program
var main = function(greet) {
  var hello = "hello " + greet;
  return hello;
};
console.log(main('world'));

>>> # minify print
>>> print(es5.minify_print(program_source, obfuscate=True))
var main=function(b){var a="hello "+b;return a;};console.log(main('world'));

>>> import json
>>> from io import StringIO
>>> from calmjs.parse.unparsers.es5 import pretty_printer
>>> from calmjs.parse.sourcemap import encode_sourcemap, write
>>> stream_p = StringIO()
>>> print_p = pretty_printer()
>>> rawmap_p, _, names_p = write(print_p(program), stream_p)
>>> sourcemap_p = encode_sourcemap(
...     'demo.min.js', rawmap_p, ['custom_name.js'], names_p)
>>> print(json.dumps(sourcemap_p, indent=2, sort_keys=True))
{
  "file": "demo.min.js",
  "mappings": "AAEA;IACI;IACA;AACJ;AACA;",
  "names": [],
  "sources": [
    "custom_name.js"
  ],
  "version": 3
}
>>> print(stream_p.getvalue())
var main = function(greet) {
...

>>> from calmjs.parse.unparsers.es5 import minify_printer
>>> program.sourcepath = 'demo.js'  # say this was opened there
>>> stream_m = StringIO()
>>> print_m = minify_printer(obfuscate=True, obfuscate_globals=True)
>>> sourcemap_m = encode_sourcemap(
...     'demo.min.js', *write(print_m(program), stream_m))
>>> print(json.dumps(sourcemap_m, indent=2, sort_keys=True))
{
  "file": "demo.min.js",
  "mappings": "AAEA,IAAIA,CAAK,CAAE,SAASC,CAAK,CAAE,CACvB,...,YAAYF,CAAI",
  "names": [
    "main",
    "greet",
    "hello"
  ],
  "sources": [
    "demo.js"
  ],
  "version": 3
}
>>> print(stream_m.getvalue())
var a=function(b){var a="hello "+b;return a;};console.log(a('world'));

>>> from calmjs.parse import io
>>> h4_program_src = open('/tmp/html4.js')
>>> h4_program_min = open('/tmp/html4.min.js', 'w+')
>>> h4_program_map = open('/tmp/html4.min.js.map', 'w+')
>>> h4_program = io.read(es5, h4_program_src)
>>> print(h4_program)
var bold = function(s) {
  return '<b>' + s + '</b>';
};
var italics = function(s) {
  return '<i>' + s + '</i>';
};
>>> io.write(print_m, h4_program, h4_program_min, h4_program_map)
>>> pos = h4_program_map.seek(0)
>>> print(h4_program_map.read())
{"file": "html4.min.js", "mappings": ..., "version": 3}
>>> pos = h4_program_min.seek(0)
>>> print(h4_program_min.read())
var b=function(a){return'<b>'+a+'</b>';};var a=function(a){...};
//# sourceMappingURL=html4.min.js.map

>>> files = [open('/tmp/html4.js'), open('/tmp/legacy.js')]
>>> combined = open('/tmp/combined.js', 'w+')
>>> io.write(print_p, (io.read(es5, f) for f in files), combined, combined)
>>> pos = combined.seek(0)
>>> print(combined.read())
var bold = function(s) {
    return '<b>' + s + '</b>';
};
var italics = function(s) {
    return '<i>' + s + '</i>';
};
var marquee = function(s) {
    return '<marquee>' + s + '</marquee>';
};
var blink = function(s) {
    return '<blink>' + s + '</blink>';
};
//# sourceMappingURL=data:application/json;base64;...

Advanced usage

>>> from calmjs.parse.unparsers.es5 import Unparser
>>> from calmjs.parse.rules import indent
>>> from calmjs.parse.rules import obfuscate
>>> pretty_obfuscate = Unparser(rules=(
...     # note that indent must come after, so that the whitespace
...     # handling rules by indent will shadow over the minimum set
...     # provided by obfuscate.
...     obfuscate(obfuscate_globals=False),
...     indent(indent_str='    '),
... ))
>>> math_module = es5(u'''
... (function(root) {
...   var fibonacci = function(count) {
...     if (count < 2)
...       return count;
...     else
...       return fibonacci(count - 1) + fibonacci(count - 2);
...   };
...
...   var factorial = function(n) {
...     if (n < 1)
...       throw new Error('factorial where n < 1 not supported');
...     else if (n == 1)
...       return 1;
...     else
...       return n * factorial(n - 1);
...   }
...
...   root.fibonacci = fibonacci;
...   root.factorial = factorial;
... })(window);
...
... var value = window.factorial(5) / window.fibonacci(5);
... console.log('the value is ' + value);
... ''')
>>> print(''.join(c.text for c in pretty_obfuscate(math_module)))
(function(b) {
    var a = function(b) {
        if (b < 2) return b;
        else return a(b - 1) + a(b - 2);
    };
    var c = function(a) {
        if (a < 1) throw new Error('factorial where n < 1 not supported');
        else if (a == 1) return 1;
        else return a * c(a - 1);
    };
    b.fibonacci = a;
    b.factorial = c;
})(window);
var value = window.factorial(5) / window.fibonacci(5);
console.log('the value is ' + value);

>>> from calmjs.parse import es5
>>> from calmjs.parse.asttypes import Object, VarDecl, FunctionCall
>>> from calmjs.parse.walkers import Walker
>>> walker = Walker()
>>> declarations = es5(u'''
... var i = 1;
... var s = {
...     a: "test",
...     o: {
...         v: "value"
...     }
... };
... foo({foo: "bar"});
... function bar() {
...     var t = {
...         foo: "bar",
...     };
...     return t;
... }
... foo.bar = bar;
... foo.bar();
... ''')
>>> # print out the object nodes that were part of some assignments
>>> for node in walker.filter(declarations, lambda node: (
...         isinstance(node, VarDecl) and
...         isinstance(node.initializer, Object))):
...     print(node.initializer)
...
{
  a: "test",
  o: {
    v: "value"
  }
}
{
  foo: "bar"
}
>>> # print out all function calls
>>> for node in walker.filter(declarations, lambda node: (
...         isinstance(node, FunctionCall))):
...     print(node.identifier)
...
foo
foo.bar

Limitations

Comments currently may be incomplete

Due to the implementation of the lexer/parser along with how the ast node types have been implemented, there are restrictions on where the comments may be exposed if enabled. Currently, such limitations exists for nodes that are created by production rules that consume multiple lexer tokens at once - only comments preceding the first token will be captured, with all remaining comments discarded.

For example, this limitation means that any comments before the else token will be omitted (as the comment will be provided by the if token), as the production rule for an If node consumes both these tokens and the node as implemented only provides a single slot for comments. Likewise, any comments before the : token in a ternary statement will also be discarded as that is the second token consumed by the production rule that produces a Conditional node.

Troubleshooting

>>> from calmjs.parse.parsers import es5
>>> parser = es5.Parser()
Traceback (most recent call last):
  ...
  File "c:\python35\....\ply\lex.py", line 1043, in lex
    lexobj.writetab(lextab, outputdir)
  File "c:\python35\....\ply\lex.py", line 195, in writetab
    tf.write('_lexstatere   = %s\n' % repr(tabre))
  File "c:\python35\lib\encodings\cp1252.py", line 19, in encode
    return codecs.charmap_encode(input,self.errors,encoding_table)[0]
UnicodeEncodeError: 'charmap' codec can't encode character '\u02c1' ...

$ python -m calmjs.parse.parsers.optimize

Contribute

• Issue Tracker: https://github.com/calmjs/calmjs.parse/issues

• Source Code: https://github.com/calmjs/calmjs.parse

Legal

The calmjs.parse package is copyright (c) 2017 Auckland Bioengineering Institute, University of Auckland. The calmjs.parse package is licensed under the MIT license (specifically, the Expat License), which is also the same license that the package slimit was released under.

The lexer, parser and the other types definitions portions were originally imported from the slimit package; slimit is copyright (c) Ruslan Spivak.

The Calmjs project is copyright (c) 2017 Auckland Bioengineering Institute, University of Auckland.

1.2.3 - 2020-03-17

• Correct usage of __doc__ to support level 2 optimized mode. [ #29 #30 ]

1.2.2 - 2020-01-18

• Correctly include LICENSE file in sdist. [ #27 #28 ]

• Include the correct test data general form for some previously added test cases to better accommodate already planned future features.

1.2.1 - 2019-11-21

• Fix the issue of failures with regex statement that occur due to lexer being in a state where the disambiguation between REGEX or DIV token types is not immediately possible, as tokens such as RBRACE, PLUSPLUS or MINUSMINUS must be consumed by parser in order to be disambiguated, but due to the lookahead nature done by yacc, the DIV token will be prematurely produced and the only way to achieve this is during the error handling stage. [ #25 #26 ]

• Part of the previous fix also removed newline or comment tokens from being reported as part of parsing error messages.

1.2.0 - 2019-08-15

• Partial support for parsing of comments. Currently not all comments will be captured during parsing, due to the desire to simplify access of them through the asttypes.Node instances with the generic comments attribute provided by it. [ #24 ]

  • Enabled by passing with_comments=True to the parser..

  • The limitation lines in the fact that if a node maps to multiple tokens (e.g. if...else), the comments that lie immediate before the first will be captured, while the comments that lie immediate to the subsequent ones will be omitted. The fix would involve providing a full syntax tree node types, and that the parser rules would need to be implemented in a more amenable manner such that the generation of such could be done.

  • All comments that lie immediately before the node are accessible using the comments attribute.

  • These comments nodes will not be yielded via the children() method.

  • Various features and methods have been updated to account for comments. Notably, sourcemap generation will be able to deal with source fragments that contain newlines provided that both colno and lineno are provided.

• Correctly fail on incorrect hexadecimal/unicode escape sequences while reporting the specific character location; also report on the starting position of an unterminated string literal. [ #23 ]

1.1.3 - 2018-11-08

• Correct issues with certain non-optional spaces being omitted for the minify print cases, which caused malformed outputs. [ #22 ]

1.1.2 - 2018-08-20

• Default repr on synthetic nodes or nodes without column or row number assigned should no longer error. [ #20 ]

• The same line terminator regex introduced in 1.1.0 used for line continuation in strings now applied to the line terminator pattern to the lexer, such that the line numbering is corrected for the Windows specific <CR><LF> sequence. [ #21 ]

1.1.1 - 2018-08-11

• Ensure that the accounting of layout rule chunks is done correctly in the case where layout handlers specified a tuple of layout rules for combined handling. [ #19 ]

  • The issue caused by this error manifest severely in the case where multiple layout rule tokens are produced in a manner that repeats a pattern that also have a layout handler rule for them, which does not typically happen for normal code with the standard printers (as layout chunks are many and they generally do not result in a repeated pattern that gets consumed). However this is severely manifested in the case of minified output with semicolons dropped, as that basically guarantee that any series of closing blocks that fit the pattern to be simply dropped.

1.1.0 - 2018-08-07

• Correct the implementation of line continuation in strings. This also meant a change in the minify unparser so that it will continue to remove the line continuation sequences. [ #16 ]

• Correct the implementation of ASI (automatic semicolon insertion) by introducing a dedicated token type, such that the production of empty statement can no longer happen and that distinguishes it from production of statements that should not have ASI applied, such that incorrectly successful parsing due to this issue will no longer result. [ #18 rspivak/slimit#29 rspivak/slimit#101 ]

1.0.1 - 2018-04-19

• Ensure that the es5 Unparser pass on the prewalk_hooks argument in its constructor.

• Minor packaging fixes; also include optimization modules for ply-3.11.

1.0.0 - 2017-09-26

Full support for sourcemaps; changes that make it possible follows:

• High level read/write functionality provided by a new io module.

• There is now a Deferrable rule type for marking certain Tokens that need extra handling. The support for this has changed the various API that deals with setting up of this.

• For support of the sourcemap generation, a number of new ruletypes have been added.

• The sourcemap write function had its argument order modified to better support the sourcepath tracking feature of input Nodes. Its return value also now match the ordering of the encode_sourcemap function.

• The chunk types in ruletypes have been renamed, and also a new type called StreamFragment is introduced, so that multiple sources output to a single stream can be properly tracked by the source mapping processes.

• rspivak/slimit#66 should be fully supported now.

Minify printer now has ability to shorten/obfuscate identifiers:

• Provide a name obfuscation function for shortening identifiers, to further achieve minified output. Note that this does not yet fully achieve the level of minification slimit had; future versions may implement this functionality as various AST transformations.

• Also provided ability to drop unneeded semicolons.

Other significant changes:

• Various changes to internal class and function names for the 1.0.0 release. A non exhaustive listing of changes to modules relative to the root of this package name as compared to previous major release follows:

  asttypes

  • All slimit compatibility features removed.

  • Switch (the incorrect version) removed.

  • SwitchStatement -> Switch

  • SetPropAssign constructor: parameters -> parameter

  • UnaryOp -> UnaryExpr

  • Other general deprecated features also removed.

  factory

  • Factory -> SRFactory

  visitors

  • Removed (details follow).

  walkers

  • visitors.generic.ReprVisitor -> walkers.ReprWalker

  layouts

  • Module was split and reorganised; the simple base ones can be found in handlers.core, the indentation related features are now in handlers.indentation.

  unparsers.base

  • .default_layout_handlers -> handlers.core.default_rules

  • .minimum_layout_handlers -> handlers.core.minimum_rules

  unparsers.prettyprint

  • Renamed to unparsers.walker.

  • The implementation was actually standard tree walking, no correctly implemented visitor functions/classes were ever present.

  vlq

  • .create_sourcemap -> sourcemap.create_sourcemap

• Broke up the visitors class as they weren’t really visitors as described. The new implementations (calmjs.parse-0.9.0) were really walkers, so move them to that name and leave it at that. Methods were also renamed to better reflect their implementation and purpose.

• Many slimit compatibility modules, classes and incorrectly implemented functionalities removed.

• The usage of the Python 3 str type (unicode in Python 2) is now enforced for the parser, to avoid various failure cases where mismatch types occur.

• The base Node asttype has a sourcepath attribute which is to be used for tracking the original source of the node; if assigned, all its subnodes without sourcepath defined should be treated as from that source.

• Also provide an even higher level function for usage with streams through the calmjs.parse.io module.

• Semicolons and braces added as structures to be rendered.

Bug fixes:

• Functions starting with a non-word character will now always have a whitespace rendered before it to avoid syntax error.

• Correct an incorrect iterator usage in the walk function.

• Ensure List separators don’t use the rowcol positions of a subsequent Elision node.

• Lexer will only report real lexer tokens on errors (ASI generated tokens are now dropped as they don’t exist in the original source which results in confusing rowcol reporting).

• rspivak/slimit#57, as it turns out '\0' is not considered to be octal, but is a <NUL> character, which the rule to parse was not actually included in the lexer patches that were pulled in previous to this version.

• rspivak/slimit#75, Option for shadowing of names of named closures, which is now disabled by default (obfuscated named closures will not be shadowed by other obfuscated names in children).

• Expressions can no longer contain an unnamed function.

0.10.1 - 2017-08-26

• Corrected the line number reporting for the lexer, and correct the propagation of that to the parser and the Node subclasses. Fixes the incorrect implementation added by moses-palmer/slimit@8f9a39c7769 (where the line numbers are tabulated incorrectly when comments are present, and also the yacc tracking added by moses-palmer/slimit@6aa92d68e0 (where the custom lexer class does not provide the position attributes required by ply).

• Implemented bookkeeping of column numbers.

• Made other various changes to AST but for compatibility reasons (to not force a major semver bump) they are only enabled with a flag to the ES5 parser.

• Corrected a fault with how switch/case statements are handled in a way that may break compatibility; fixes are only enabled when flagged. rspivak/slimit#94

• The repr form of Node now shows the line/col number info by default; the visit method of the ReprVisitor class have not been changed, only the invocation of it via the callable form has as that is the call target for __repr__. This is a good time to mention that named methods afford the most control for usage as documented already.

• Parsers now accept an asttypes module during its construction.

• Provide support for source map generation classes.

• Introduced a flexible visitor function/state class that accepts a definition of rules for the generation of chunk tuples that are compatible for the source map generation. A new way for pretty printing and minification can be achieved using this module.

0.9.0 - 2017-06-09

• Initial release of the fork of slimit.parser and its parent modules as calmjs.parse.

• This release brings in a number of bug fixes that were available via other forks of slimit, with modifications or even a complete revamp.

• Issues addressed includes:

  • rspivak/slimit#52, rspivak/slimit#59, rspivak/slimit#81, rspivak/slimit#90 (relating to conformance of ecma-262 7.6 identifier names)

  • rspivak/slimit#54 (fixed by tracking scope and executable current token in lexer)

  • rspivak/slimit#57, rspivak/slimit#70 (octal encoding (e.g 0), from redapple/slimit@a93204577f)

  • rspivak/slimit#62 (formalized into a unittest that passed)

  • rspivak/slimit#73 (specifically the desire for a better repr; the minifier bits are not relevant to this package)

  • rspivak/slimit#79 (tab module handling was completely reimplemented)

  • rspivak/slimit#82 (formalized into a unittest that passed)

• Include various changes gathered by rspivak/slimit#65, which may be the source of some of the fixes listed above.