dan/oxc
1
0
Fork 0
mirror of https://github.com/danbulant/oxc synced 2026-05-25 12:51:57 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions
6640 commits 1 branch 0 tags 45 MiB
Commit graph

84 commits

Author SHA1 Message Date
Boshen
413973df66 feat(codegen): print linked and external legal comment (#7059) 
part of #7050
 2024-11-02 04:08:14 +00:00
Boshen
ee27b92465 feat(codegen): print eof legal comments (#7058) 
part of https://github.com/oxc-project/oxc/issues/7050
 2024-11-01 16:56:57 +00:00
Boshen
6516f9eabc feat(codegen): print inline legal comments (#7054) 
part of https://github.com/oxc-project/oxc/issues/7050
 2024-11-01 16:56:56 +00:00
Boshen
0bb1aa4c64 refactor(codegen): move options to its own file (#7053) 2024-11-01 09:28:59 +00:00
DonIsaac
374b972143 docs(codegen): add #![warn(missing_docs)] to oxc_codegen (#6711) 
Part of https://github.com/oxc-project/backlog/issues/130
 2024-10-21 00:13:57 +00:00
Boshen
e5ed6a56a8 feat(codegen): print negative numbers (#6624) 2024-10-16 22:38:54 +00:00
overlookmotel
c0e9d7eb77 refactor(codegen)!: Codegen::into_source_text consume Codegen (#6539) 
Breaking change. `Codegen::into_source_text` consume `Codegen`, instead of taking the `CodeBuffer` and substituting an empty one.

Keeping the `Codegen` alive seems unintuitive, as its state is then out of sync. Consuming the `CodeBuffer` is also marginally cheaper.
 2024-10-14 01:09:03 +00:00
overlookmotel
77f3a1a9dd perf(codegen): check last char with byte methods (#6509) 
When checking what last char in buffer is, avoid calculating a `char` when only comparing to an ASCII char (byte) anyway.
 2024-10-14 01:09:02 +00:00
overlookmotel
51fc63da45 refactor(codegen): rename CodeBuffer::print_bytes_unchecked method (#6517) 2024-10-13 13:14:49 +00:00
overlookmotel
782f0a7a69 refactor(codegen)!: rename print_char method to print_ascii_byte (#6512) 
`print_ascii_byte` is more descriptive of what this method does, and makes clear the invariant that `byte` must be ASCII.
 2024-10-13 12:51:37 +00:00
overlookmotel
74206204dd refactor(codegen): add CodeBuffer::as_bytes method (#6498) 
`as_bytes` is clearer than `as_ref` (what kind of ref?)
 2024-10-13 10:05:10 +00:00
DonIsaac
204bf5533e refactor(codegen): add CodeBuffer to fix soundness hole (#6148) 
# What This PR Does

Adds `CodeBuffer`, a simple wrapper over a `Vec<u8>` with a protective and reduced API for upholding UTF-8 validity guarantees. Closes #6147.

Note that this struct is actually quite small. Most of the added lines are doc comments.
 2024-10-13 09:14:48 +00:00
Boshen
91c87dd380 refactor(codegen)!: remove Codegen::enableSourceMap API (#6452) 2024-10-12 04:56:43 +00:00
Boshen
7645e5c34b refactor(codegen)!: remove CommentOptions API (#6451) 2024-10-11 13:53:28 +00:00
Boshen
520096030a refactor(oxc)!: remove passing Trivias around (#6446) 
part of #6426
 2024-10-11 06:09:25 +00:00
Boshen
020bb80b65 refactor(codegen)!: change to CodegenReturn::code and CodegenReturn::map (#6310) 2024-10-06 05:05:47 +00:00
Dunqing
cca433f84a feat(codegen): print vite / webpack special comments (#6021) 
Related: https://github.com/oxc-project/oxc/issues/1046#issuecomment-2347345319
Close: #6024
 2024-09-26 15:19:31 +00:00
Dunqing
fe696f0dec refactor(codegen): simplify printing annotation comments (#6027) 
Simplify printing annotation comments, reusing the architecture of printing JSDoc.
 2024-09-25 06:27:18 +00:00
Boshen
bcdbba3981 feat(codegen): print jsdoc comments that are attached to statements and class elements (#5845) 
I am unable to print all comments correctly. Comments have way too much semantic meaning in JavaScript.

This PR reduces the scope to only print jsdoc comments that are attached to statements and class elements, in order to get isolated declarations shipped.
 2024-09-18 07:58:22 +00:00
Boshen
42dcadfccf fix(parser): hashbang comment should not keep the end newline char (#5844) 
Previously it included a newline in the value

```
  "hashbang": {
    "type": "Hashbang",
    "start": 0,
    "end": 16,
    "value": "/usr/bin/node\n"
  },
```

This change will also make the lexer emit a `\n` token, which will make comment position detection correct.
 2024-09-18 05:41:44 +00:00
DonIsaac
26386da01e feat(codegen): have with_source_text reserve memory for code buffer (#5823) 
`Codegen::with_source_text` gives us enough information to use `Codegen::with_capacity`. This change makes the API cleaner as users have to provide less redundant information.
 2024-09-18 03:27:43 +00:00
Boshen
e613a3d877 refactor(codegen): prepare to add leading comments by adding a template method pattern (#5784) 2024-09-15 09:41:53 +00:00
Boshen
1bed5ce2a5 chore: run cargo +nightly fmt to sort imports (#5503) 
They are never going to be stable are they ... cedf7a4daa/.rustfmt.toml (L8-L16)
 2024-09-06 04:04:26 +00:00
rzvxa
0df1d9d97b fix(ast, codegen, linter): panics in fixers. (#5431) 
Closes #5434

https://github.com/oxc-project/oxlint-ecosystem-ci/actions/runs/10685877915/job/29619941099
 2024-09-05 05:59:22 +00:00
overlookmotel
12a7607bac perf(codegen): inline Codegen::print_list (#5221) 
Revert #5192 and add a comment that it's not a perf gain.

This was really surprising to me, but the benchmarks do demonstrate it.

Please see the benchmarks commit-by-commit on this PR. Adding `#[inline]` to the function does give +1% gain, but it's no better than it was before #5192. So I think preferable to just revert to the simpler original.

I think likely explanation is that the compiler is already performing this optimization itself. And if it does it itself, then it understands the code better, and can then make better decisions about inlining.

https://godbolt.org/z/xzhWWeMoe seems to demonstrate this - there are 2 calls to `Item::gen` in the generated assembly, so it has split the loop into 2.
 2024-08-26 10:28:49 +00:00
Boshen
fb847bd0ba perf(codegen): slightly faster print_list (#5192) 2024-08-26 01:37:48 +00:00
Boshen
543cad6370
refactor(codegen): remove some pub APIs 2024-08-25 13:32:55 +08:00
Boshen
7b1546bb66
fix(codegen): do not print comments when --minify 2024-08-24 13:49:34 +08:00
IWANABETHATGUY
b7db235065
fix: comments gen regression (#5003) 
try to fix: https://github.com/rolldown/rolldown/issues/2013
1. Before we only considering the ast is untouched, but considering the
scenario.
```js

const a = /*__PURE__*/ test(),
//    ^^^              ^^^^^^ is removed during transform
	b = a();

```
Then according to the previous algorithm, `PURE` will attach to `b =
a()`
2. Now, we try to attach comments as much as possible unless the
comments are separated by comments, for the case above, `PURE` will not
be attached to `a()` since the content between `b = a()` and `/*
__PURE__*/` is not all whitespace.
3. we added back `MoveMap`, for the special case 
```js
/*__NODE_SIDE_EFFECTS__*/ export const c = 100;
// ^^^^^^^^^^^^^^^^^^^^^         should be attached to first declarator,
//                        ^^^^^^ are not whitespace

```
 2024-08-20 18:57:12 +08:00
Boshen
cd9cf5efd8
refactor(oxc): remove remove_whitespace 2024-08-20 18:27:59 +08:00
Boshen
ce4d4698b4 feat(codegen)!: remove const generic MINIFY (#5001) 
This is a premature optimization, makes the code complicated, and bloats the final binary size.

The minify option is moved to `CodegenOptions`
 2024-08-20 08:13:27 +00:00
IWANABETHATGUY
d49fb160e3 feat(oxc_codegen): support generate range leading comments (#4898) 
1. Support print range leading comments for specific expr or stmt
 2024-08-15 04:20:48 +00:00
IWANABETHATGUY
1808529a0c
fix(codegen): dedupe pure annotation comments (#4862) 
Close https://github.com/oxc-project/oxc/issues/4843
 2024-08-13 13:53:28 +08:00
Boshen
a558492bf9 feat(codegen): implement BinaryExpressionVisitor (#4548) 
part of https://github.com/oxc-project/backlog/issues/58

`monitor-oxc` run: https://github.com/oxc-project/monitor-oxc/actions/runs/10179047831
binary expression stack length tally using `counts` in top 100 npm packages from monitor-oxc:

```
29772 counts
(  1)    17652 (59.3%, 59.3%): 0
(  2)     5772 (19.4%, 78.7%): 1
(  3)     3204 (10.8%, 89.4%): 2
(  4)     1276 ( 4.3%, 93.7%): 3
(  5)      616 ( 2.1%, 95.8%): 4
(  6)      308 ( 1.0%, 96.8%): 5
(  7)      202 ( 0.7%, 97.5%): 6
(  8)      168 ( 0.6%, 98.1%): 7
(  9)      114 ( 0.4%, 98.5%): 9
( 10)       90 ( 0.3%, 98.8%): 8
( 11)       84 ( 0.3%, 99.0%): 13
( 12)       58 ( 0.2%, 99.2%): 10
( 13)       48 ( 0.2%, 99.4%): 12
( 14)       32 ( 0.1%, 99.5%): 11
( 15)       20 ( 0.1%, 99.6%): 134
( 16)       16 ( 0.1%, 99.6%): 18
( 17)       16 ( 0.1%, 99.7%): 20
( 18)       12 ( 0.0%, 99.7%): 19
( 19)       12 ( 0.0%, 99.8%): 35
( 20)       12 ( 0.0%, 99.8%): 51
( 21)       10 ( 0.0%, 99.8%): 15
( 22)        6 ( 0.0%, 99.9%): 17
( 23)        6 ( 0.0%, 99.9%): 21
( 24)        6 ( 0.0%, 99.9%): 45
( 25)        4 ( 0.0%, 99.9%): 14
( 26)        4 ( 0.0%, 99.9%): 26
( 27)        4 ( 0.0%, 99.9%): 53
( 28)        2 ( 0.0%, 99.9%): 172
( 29)        2 ( 0.0%, 99.9%): 214
( 30)        2 ( 0.0%,100.0%): 22
( 31)        2 ( 0.0%,100.0%): 27
( 32)        2 ( 0.0%,100.0%): 28
( 33)        2 ( 0.0%,100.0%): 29
( 34)        2 ( 0.0%,100.0%): 31
( 35)        2 ( 0.0%,100.0%): 36
( 36)        2 ( 0.0%,100.0%): 46
( 37)        2 ( 0.0%,100.0%): 55
```
 2024-07-31 12:44:19 +00:00
Boshen
35654e665c feat(codegen): align operator precedence with esbuild (#4509) 
closes #4339
 2024-07-28 11:48:51 +00:00
Boshen
6a94e3f573 fix(codegen): fixes for esbuild test cases (#4503) 2024-07-28 08:57:15 +00:00
Boshen
3d88f20cbb
fix(codegen): print shorthand for all { x } variants (#4374) 
closes #4340
 2024-07-21 19:54:21 +08:00
Boshen
a71787572e
chore: remove unsafe_code = "warn" rust lint 
Feels too verbose as we already have unsafe comment turned on
 2024-07-15 10:39:08 +08:00
Boshen
d1c4be0020
refactor(codegen): clean up annotation_comment 2024-07-14 15:55:50 +08:00
Boshen
e167ef79c6 fix(codegen): print parenthesis properly (#4245) 
`TSParenthesizedType` handles parenthesis in ts types.

It should be considered a bug if parenthesis is not printed correctly after this PR.
 2024-07-14 04:13:10 +00:00
IWANABETHATGUY
66b455a2b2
fix(oxc_codegen): avoid print same pure comments multiple time (#4230) 
## Before 
```bash
Original:
const builtInSymbols = new Set(
  /*#__PURE__*/
  Object.getOwnPropertyNames(Symbol)
    .filter(key => key !== 'arguments' && key !== 'caller')
)


Printed:
const builtInSymbols = new Set(/*#__PURE__*/ /*#__PURE__*/ Object.getOwnPropertyNames(Symbol).filter((key) => key !== "arguments" && key !== "caller"));

Minified:
const builtInSymbols=new Set(Object.getOwnPropertyNames(Symbol).filter((key)=>key!=="arguments"&&key!=="caller"))

```

## After
```bash
Original:
const builtInSymbols = new Set(
  /*#__PURE__*/
  Object.getOwnPropertyNames(Symbol)
    .filter(key => key !== 'arguments' && key !== 'caller')
)


Printed:
const builtInSymbols = new Set(/*#__PURE__*/  Object.getOwnPropertyNames(Symbol).filter((key) => key !== "arguments" && key !== "caller"));

Minified:
const builtInSymbols=new Set(Object.getOwnPropertyNames(Symbol).filter((key)=>key!=="arguments"&&key!=="caller"))

```
 2024-07-12 23:07:43 +08:00
Boshen
83c2c62f7b feat(codegen): add option for choosing quotes; remove slow choose_quot method (#4219) 2024-07-12 03:08:22 +00:00
Boshen
e3e663bae4 feat(mangler): initialize crate and integrate into minifier (#4197) 2024-07-11 10:35:13 +00:00
Boshen
aa22073736 refactor(codegen): improve print API (#4196) 2024-07-11 08:41:04 +00:00
Luca Bruno
5731e3957f
refactor(ast)!: store span details inside comment struct (#4132) 
This tweaks `Comment` definition in order to internally store the start
and end position of its span.

Closes: https://github.com/oxc-project/oxc/issues/4069
 2024-07-09 23:23:43 +08:00
IWANABETHATGUY
365d9ba252
feat(oxc_codegen): generate annotation comments before CallExpression and NewExpression (#4119) 
1. test case copy from `vue/core`, here are all usages of `#__PURE__` in
`vue/core`
https://gist.github.com/IWANABETHATGUY/c7911ecd98467a2969b2a994a34d32bc#file-pure_annotation_in_vue_repo-sh
2. Also took a look in other codebase,
https://github.com/search?q=%23__PURE__&type=code, most of the usage of
`#__PURE__` attached as leading comment before `CallExpression` and
`NewExpression`
 2024-07-09 00:22:28 +08:00
Boshen
5472b7c990
fix(codegen): 256 indentations level is not enough for codegen 2024-07-08 00:25:37 +08:00
Boshen
aaac2d8775
fix(codegen): preserve parentheses from AST instead calculating from operator precedence (#4055) 
…operator precedence

Calculating from operator precedence is currently unsafe and will result
incorrect semantics.
 2024-07-05 14:01:17 +08:00
overlookmotel
4f7ff7e3ad perf: do not pass &Atom to functions (#3818) 
`Atom` is just a wrapper around `&str`, so better not to pass `&Atom` to functions, as that's a double-reference. Prefer `Atom` or `&str` instead to avoid indirection.
 2024-06-22 04:48:00 +00:00
Boshen
051ceb6539
chore: improve some format by running cargo +nightly fmt 2024-06-19 00:48:30 +08:00