JoBoEVD/UglifyJS
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

make comments output more robust (#2633)

Browse Source 
- improve handling of comments right after `return`
- retain comments after `OutputStream`
- preserve trailing comments
- fix handling of new line before comments
- handle comments around parentheses

fixes #88
fixes #112
fixes #218
fixes #372
fixes #2629
This commit is contained in:
Alex Lam S.L
2017-12-22 04:59:54 +08:00
committed by GitHub
parent 4113609dd4
commit edb4e3bd52
7 changed files with 401 additions and 90 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
lib/ast.js
View File

@@ -87,7 +87,7 @@ function DEFNODE(type, props, methods, base) {
return ctor;
};
var AST_Token = DEFNODE("Token", "type value line col pos endline endcol endpos nlb comments_before file raw", {
var AST_Token = DEFNODE("Token", "type value line col pos endline endcol endpos nlb comments_before comments_after file raw", {
}, null);
var AST_Node = DEFNODE("Node", "start end", {

212
lib/output.js
View File

@@ -52,6 +52,7 @@ function is_some_comments(comment) {
function OutputStream(options) {
var readonly = !options;
options = defaults(options, {
ascii_only : false,
beautify : false,
@@ -199,6 +200,7 @@ function OutputStream(options) {
var might_need_space = false;
var might_need_semicolon = false;
var might_add_newline = 0;
var need_newline_indented = false;
var last = "";
var mapping_token, mapping_name, mappings = options.source_map && [];
@@ -257,6 +259,13 @@ function OutputStream(options) {
function print(str) {
str = String(str);
var ch = str.charAt(0);
if (need_newline_indented && ch) {
need_newline_indented = false;
if (ch != "\n") {
print("\n");
indent();
}
}
var prev = last.charAt(last.length - 1);
if (might_need_semicolon) {
might_need_semicolon = false;
@@ -427,6 +436,109 @@ function OutputStream(options) {
return OUTPUT;
};
function prepend_comments(node) {
var self = this;
var start = node.start;
if (!(start.comments_before && start.comments_before._dumped === self)) {
var comments = start.comments_before;
if (!comments) {
comments = start.comments_before = [];
}
comments._dumped = self;
if (node instanceof AST_Exit && node.value) {
var tw = new TreeWalker(function(node) {
var parent = tw.parent();
if (parent instanceof AST_Exit
|| parent instanceof AST_Binary && parent.left === node
|| parent.TYPE == "Call" && parent.expression === node
|| parent instanceof AST_Conditional && parent.condition === node
|| parent instanceof AST_Dot && parent.expression === node
|| parent instanceof AST_Sequence && parent.expressions[0] === node
|| parent instanceof AST_Sub && parent.expression === node
|| parent instanceof AST_UnaryPostfix) {
var text = node.start.comments_before;
if (text && text._dumped !== self) {
text._dumped = self;
comments = comments.concat(text);
}
} else {
return true;
}
});
tw.push(node);
node.value.walk(tw);
}
if (current_pos == 0) {
if (comments.length > 0 && options.shebang && comments[0].type == "comment5") {
print("#!" + comments.shift().value + "\n");
indent();
}
var preamble = options.preamble;
if (preamble) {
print(preamble.replace(/\r\n?|[\n\u2028\u2029]|\s*$/g, "\n"));
}
}
comments = comments.filter(comment_filter, node);
if (comments.length == 0) return;
var last_nlb = /(^|\n) *$/.test(OUTPUT);
comments.forEach(function(c, i) {
if (!last_nlb) {
if (c.nlb) {
print("\n");
indent();
last_nlb = true;
} else if (i > 0) {
space();
}
}
if (/comment[134]/.test(c.type)) {
print("//" + c.value + "\n");
indent();
last_nlb = true;
} else if (c.type == "comment2") {
print("/*" + c.value + "*/");
last_nlb = false;
}
});
if (!last_nlb) {
if (start.nlb) {
print("\n");
indent();
} else {
space();
}
}
}
}
function append_comments(node, tail) {
var self = this;
var token = node.end;
if (!token) return;
var comments = token[tail ? "comments_before" : "comments_after"];
if (comments && comments._dumped !== self) {
comments._dumped = self;
comments.filter(comment_filter, node).forEach(function(c, i) {
if (need_newline_indented || c.nlb) {
print("\n");
indent();
need_newline_indented = false;
} else if (i > 0 || !tail) {
space();
}
if (/comment[134]/.test(c.type)) {
print("//" + c.value);
need_newline_indented = true;
} else if (c.type == "comment2") {
print("/*" + c.value + "*/");
}
});
}
}
var stack = [];
return {
get : get,
@@ -464,7 +576,8 @@ function OutputStream(options) {
with_square : with_square,
add_mapping : add_mapping,
option : function(opt) { return options[opt] },
comment_filter : comment_filter,
prepend_comments: readonly ? noop : prepend_comments,
append_comments : readonly ? noop : append_comments,
line : function() { return current_line },
col : function() { return current_col },
pos : function() { return current_pos },
@@ -500,9 +613,10 @@ function OutputStream(options) {
use_asm = active_scope;
}
function doit() {
self.add_comments(stream);
stream.prepend_comments(self);
self.add_source_map(stream);
generator(self, stream);
stream.append_comments(self);
}
stream.push_node(self);
if (force_parens || self.needs_parens(stream)) {
@@ -519,77 +633,10 @@ function OutputStream(options) {
AST_Node.DEFMETHOD("print_to_string", function(options){
var s = OutputStream(options);
if (!options) s._readonly = true;
this.print(s);
return s.get();
});
/* -----[ comments ]----- */
AST_Node.DEFMETHOD("add_comments", function(output){
if (output._readonly) return;
var self = this;
var start = self.start;
if (start && !start._comments_dumped) {
start._comments_dumped = true;
var comments = start.comments_before || [];
// XXX: ugly fix for https://github.com/mishoo/UglifyJS2/issues/112
// and https://github.com/mishoo/UglifyJS2/issues/372
if (self instanceof AST_Exit && self.value) {
self.value.walk(new TreeWalker(function(node){
if (node.start && node.start.comments_before) {
comments = comments.concat(node.start.comments_before);
node.start.comments_before = [];
}
if (node instanceof AST_Function ||
node instanceof AST_Array ||
node instanceof AST_Object)
{
return true; // don't go inside.
}
}));
}
if (output.pos() == 0) {
if (comments.length > 0 && output.option("shebang") && comments[0].type == "comment5") {
output.print("#!" + comments.shift().value + "\n");
output.indent();
}
var preamble = output.option("preamble");
if (preamble) {
output.print(preamble.replace(/\r\n?|[\n\u2028\u2029]|\s*$/g, "\n"));
}
}
comments = comments.filter(output.comment_filter, self);
// Keep single line comments after nlb, after nlb
if (!output.option("beautify") && comments.length > 0 &&
/comment[134]/.test(comments[0].type) &&
output.col() !== 0 && comments[0].nlb)
{
output.print("\n");
}
comments.forEach(function(c){
if (/comment[134]/.test(c.type)) {
output.print("//" + c.value + "\n");
output.indent();
}
else if (c.type == "comment2") {
output.print("/*" + c.value + "*/");
if (start.nlb) {
output.print("\n");
output.indent();
} else {
output.space();
}
}
});
}
});
/* -----[ PARENTHESES ]----- */
function PARENS(nodetype, func) {
@@ -811,14 +858,21 @@ function OutputStream(options) {
self.body.print(output);
output.semicolon();
});
function print_bracketed(body, output, allow_directives) {
if (body.length > 0) output.with_block(function(){
display_body(body, false, output, allow_directives);
});
else output.print("{}");
function print_bracketed(self, output, allow_directives) {
if (self.body.length > 0) {
output.with_block(function() {
display_body(self.body, false, output, allow_directives);
});
} else {
output.print("{");
output.with_indent(output.next_indent(), function() {
output.append_comments(self, true);
});
output.print("}");
}
};
DEFPRINT(AST_BlockStatement, function(self, output){
print_bracketed(self.body, output);
print_bracketed(self, output);
});
DEFPRINT(AST_EmptyStatement, function(self, output){
output.semicolon();
@@ -913,7 +967,7 @@ function OutputStream(options) {
});
});
output.space();
print_bracketed(self.body, output, true);
print_bracketed(self, output, true);
});
DEFPRINT(AST_Lambda, function(self, output){
self._do_print(output);
@@ -1044,7 +1098,7 @@ function OutputStream(options) {
DEFPRINT(AST_Try, function(self, output){
output.print("try");
output.space();
print_bracketed(self.body, output);
print_bracketed(self, output);
if (self.bcatch) {
output.space();
self.bcatch.print(output);
@@ -1061,12 +1115,12 @@ function OutputStream(options) {
self.argname.print(output);
});
output.space();
print_bracketed(self.body, output);
print_bracketed(self, output);
});
DEFPRINT(AST_Finally, function(self, output){
output.print("finally");
output.space();
print_bracketed(self.body, output);
print_bracketed(self, output);
});
/* -----[ var/const ]----- */

15
lib/parse.js
View File

@@ -317,11 +317,7 @@ function tokenizer($TEXT, filename, html5_comments, shebang) {
}
if (!is_comment) {
ret.comments_before = S.comments_before;
S.comments_before = [];
// make note of any newlines in the comments that came before
for (var i = 0, len = ret.comments_before.length; i < len; i++) {
ret.nlb = ret.nlb || ret.comments_before[i].nlb;
}
ret.comments_after = S.comments_before = [];
}
S.newline_before = false;
return new AST_Token(ret);
@@ -1280,9 +1276,16 @@ function parse($TEXT, options) {
case "(":
next();
var ex = expression(true);
[].push.apply(start.comments_before, ex.start.comments_before);
ex.start.comments_before = start.comments_before;
start.comments_after = ex.start.comments_after;
ex.start = start;
ex.end = S.token;
expect(")");
var end = prev();
end.comments_before = ex.end.comments_before;
[].push.apply(ex.end.comments_after, end.comments_after);
end.comments_after = ex.end.comments_after;
ex.end = end;
return subscripts(ex, allow_calls);
case "[":
return subscripts(array_(), allow_calls);

2
lib/utils.js
View File

@@ -324,7 +324,7 @@ function first_in_statement(stack) {
if (p instanceof AST_Statement && p.body === node)
return true;
if ((p instanceof AST_Sequence && p.expressions[0] === node) ||
(p instanceof AST_Call && p.expression === node && !(p instanceof AST_New) ) ||
(p.TYPE == "Call" && p.expression === node ) ||
(p instanceof AST_Dot && p.expression === node ) ||
(p instanceof AST_Sub && p.expression === node ) ||
(p instanceof AST_Conditional && p.condition === node ) ||

82
test/compress/pure_funcs.js
View File

@@ -293,3 +293,85 @@ unary: {
bar();
}
}
issue_2629_1: {
options = {
side_effects: true,
}
input: {
/*@__PURE__*/ a();
/*@__PURE__*/ (b());
(/*@__PURE__*/ c)();
(/*@__PURE__*/ d());
}
expect: {}
}
issue_2629_2: {
options = {
side_effects: true,
}
input: {
/*@__PURE__*/ a(1)(2)(3);
/*@__PURE__*/ (b(1))(2)(3);
/*@__PURE__*/ (c(1)(2))(3);
/*@__PURE__*/ (d(1)(2)(3));
(/*@__PURE__*/ e)(1)(2)(3);
(/*@__PURE__*/ f(1))(2)(3);
(/*@__PURE__*/ g(1)(2))(3);
(/*@__PURE__*/ h(1)(2)(3));
}
expect: {}
}
issue_2629_3: {
options = {
side_effects: true,
}
input: {
/*@__PURE__*/ a.x(1).y(2).z(3);
/*@__PURE__*/ (a.x)(1).y(2).z(3);
/*@__PURE__*/ (a.x(1)).y(2).z(3);
/*@__PURE__*/ (a.x(1).y)(2).z(3);
/*@__PURE__*/ (a.x(1).y(2)).z(3);
/*@__PURE__*/ (a.x(1).y(2).z)(3);
/*@__PURE__*/ (a.x(1).y(2).z(3));
(/*@__PURE__*/ a).x(1).y(2).z(3);
(/*@__PURE__*/ a.x)(1).y(2).z(3);
(/*@__PURE__*/ a.x(1)).y(2).z(3);
(/*@__PURE__*/ a.x(1).y)(2).z(3);
(/*@__PURE__*/ a.x(1).y(2)).z(3);
(/*@__PURE__*/ a.x(1).y(2).z)(3);
(/*@__PURE__*/ a.x(1).y(2).z(3));
}
expect: {}
}
issue_2629_4: {
options = {
side_effects: true,
}
input: {
(/*@__PURE__*/ x(), y());
(w(), /*@__PURE__*/ x(), y());
}
expect: {
y();
w(), y();
}
}
issue_2629_5: {
options = {
side_effects: true,
}
input: {
[ /*@__PURE__*/ x() ];
[ /*@__PURE__*/ x(), y() ];
[ w(), /*@__PURE__*/ x(), y() ];
}
expect: {
y();
w(), y();
}
}

6
test/mocha/comment-filter.js
View File

@@ -14,7 +14,7 @@ describe("comment filters", function() {
it("Should be able to filter commments with the 'some' option", function() {
var ast = UglifyJS.parse("// foo\n/*@preserve*/\n// bar\n/*@license*/\n//@license with the wrong comment type\n/*@cc_on something*/");
assert.strictEqual(ast.print_to_string({comments: "some"}), "/*@preserve*/\n/*@license*/\n/*@cc_on something*/\n");
assert.strictEqual(ast.print_to_string({comments: "some"}), "/*@preserve*/\n/*@license*/\n/*@cc_on something*/");
});
it("Should be able to filter comments by passing a function", function() {
@@ -55,12 +55,12 @@ describe("comment filters", function() {
return true;
};
assert.strictEqual(ast.print_to_string({comments: f}), "#!Random comment\n//test1\n/*test2*/\n");
assert.strictEqual(ast.print_to_string({comments: f}), "#!Random comment\n//test1\n/*test2*/");
});
it("Should never be able to filter comment5 when using 'some' as filter", function() {
var ast = UglifyJS.parse("#!foo\n//foo\n/*@preserve*/\n/* please hide me */");
assert.strictEqual(ast.print_to_string({comments: "some"}), "#!foo\n/*@preserve*/\n");
assert.strictEqual(ast.print_to_string({comments: "some"}), "#!foo\n/*@preserve*/");
});
it("Should have no problem on multiple calls", function() {

172
test/mocha/comment.js
View File

@@ -47,4 +47,176 @@ describe("Comment", function() {
}, fail, tests[i]);
}
});
it("Should handle comment within return correctly", function() {
var result = uglify.minify([
"function unequal(x, y) {",
" return (",
" // Either one",
" x < y",
" ||",
" y < x",
" );",
"}",
].join("\n"), {
compress: false,
mangle: false,
output: {
beautify: true,
comments: "all",
},
});
if (result.error) throw result.error;
assert.strictEqual(result.code, [
"function unequal(x, y) {",
" // Either one",
" return x < y || y < x;",
"}",
].join("\n"));
});
it("Should handle comment folded into return correctly", function() {
var result = uglify.minify([
"function f() {",
" /* boo */ x();",
" return y();",
"}",
].join("\n"), {
mangle: false,
output: {
beautify: true,
comments: "all",
},
});
if (result.error) throw result.error;
assert.strictEqual(result.code, [
"function f() {",
" /* boo */",
" return x(), y();",
"}",
].join("\n"));
});
it("Should not drop comments after first OutputStream", function() {
var code = "/* boo */\nx();";
var ast = uglify.parse(code);
var out1 = uglify.OutputStream({
beautify: true,
comments: "all",
});
ast.print(out1);
var out2 = uglify.OutputStream({
beautify: true,
comments: "all",
});
ast.print(out2);
assert.strictEqual(out1.get(), code);
assert.strictEqual(out2.get(), out1.get());
});
it("Should retain trailing comments", function() {
var code = [
"if (foo /* lost comment */ && bar /* lost comment */) {",
" // this one is kept",
" {/* lost comment */}",
" !function() {",
" // lost comment",
" }();",
" function baz() {/* lost comment */}",
" // lost comment",
"}",
"// comments right before EOF are lost as well",
].join("\n");
var result = uglify.minify(code, {
compress: false,
mangle: false,
output: {
beautify: true,
comments: "all",
},
});
if (result.error) throw result.error;
assert.strictEqual(result.code, code);
});
it("Should correctly preserve new lines around comments", function() {
var tests = [
[
"// foo",
"// bar",
"x();",
].join("\n"),
[
"// foo",
"/* bar */",
"x();",
].join("\n"),
[
"// foo",
"/* bar */ x();",
].join("\n"),
[
"/* foo */",
"// bar",
"x();",
].join("\n"),
[
"/* foo */ // bar",
"x();",
].join("\n"),
[
"/* foo */",
"/* bar */",
"x();",
].join("\n"),
[
"/* foo */",
"/* bar */ x();",
].join("\n"),
[
"/* foo */ /* bar */",
"x();",
].join("\n"),
"/* foo */ /* bar */ x();",
].forEach(function(code) {
var result = uglify.minify(code, {
compress: false,
mangle: false,
output: {
beautify: true,
comments: "all",
},
});
if (result.error) throw result.error;
assert.strictEqual(result.code, code);
});
});
it("Should preserve new line before comment without beautify", function() {
var code = [
"function f(){",
"/* foo */bar()}",
].join("\n");
var result = uglify.minify(code, {
compress: false,
mangle: false,
output: {
comments: "all",
},
});
if (result.error) throw result.error;
assert.strictEqual(result.code, code);
});
it("Should preserve comments around IIFE", function() {
var result = uglify.minify("/*a*/(/*b*/function(){/*c*/}/*d*/)/*e*/();", {
compress: false,
mangle: false,
output: {
comments: "all",
},
});
if (result.error) throw result.error;
assert.strictEqual(result.code, "/*a*/ /*b*/(function(){/*c*/}/*d*/ /*e*/)();");
});
});