Skip to content

Replace stripDefaultUndefined with Terser --compress - #27926

Merged
sbc100 merged 1 commit into
emscripten-core:mainfrom
sbc100:terser-compress
Oct 9, 2026
Merged

Replace stripDefaultUndefined with Terser --compress#27926
sbc100 merged 1 commit into
emscripten-core:mainfrom
sbc100:terser-compress

Conversation

@sbc100

@sbc100 sbc100 commented Oct 8, 2026 •
edited
Loading

Copy link
Copy Markdown
Collaborator

Instead of a dedicated stripDefaultUndefined Acorn AST pass (added in
compression ({defaults: false, evaluate: true, keep_fargs: false})
when minifying JS without Closure Compiler. In addition to stripping
= undefined default parameter values, this performs lightweight Terser
cleanups (undefined -> void 0, single-statement block brace
removal, and constant folding).

In general, this prevents us from needing to replicate compression
passes that already exist in upstream Terser (see #27871).

Timing highlights (test/hello_world.c with -sINCLUDE_FULL_LIBRARY,
~252 KB minified JS):

  • minify_sync with compress: false: 105.5 ms
  • minify_sync with compress: {defaults: false, evaluate: true, keep_fargs: false}: 188.3 ms (+82.8 ms, or ~+63 ms net after
    removing the stripDefaultUndefined AST walk), saving ~1.4 KB.
  • On a standard -O2 hello_world.c build (~8.5 KB JS), minify_sync
    goes from 3.2 ms to 5.5 ms (+2.3 ms).

This change, in general, should prevent us from being tempted to
replicate compression passes that already exist in upstream terser
(See #27871, for example).

See: #27839

@sbc100
sbc100 requested a review from kripken October 8, 2026 15:37
sbc100
sbc100 commented Oct 8, 2026
View reviewed changes
Comment thread test/codesize/test_codesize_file_preload.expected.js
@sbc100
sbc100 requested a review from dschuff October 8, 2026 17:39
sbc100
sbc100 commented Oct 8, 2026
View reviewed changes
Comment thread test/test_other.py
kripken
kripken reviewed Oct 8, 2026
View reviewed changes

@kripken kripken left a comment

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Nice!

This looks right to me. Maybe the only question I have is if we want to run --compress in -O2 or just -O3 (I would lean towards not in -O2).

Comment thread tools/building.py
@sbc100 sbc100 mentioned this pull request Oct 8, 2026
Merged
@sbc100
sbc100 force-pushed the terser-compress branch 2 times, most recently from e8578dc to 2a0a2d8 Compare October 8, 2026 22:04
sbc100 added a commit that referenced this pull request Oct 8, 2026
@sbc100
Consolidate acorn_optimizer calls in minify_wasm_js. NFC (#27928)
e1413c3 
Previously, at `-O3`/`-Os`/`-Oz` (`expensive_optimizations`),
`minify_wasm_js` invoked `acorn-optimizer.mjs` up to four separate times
to parse and reprint JS:
1. Pre-`metadce` (`AJSDCE`, `stripDefaultUndefined`,
`--minify-whitespace`)
2. Inside `metadce` (`applyDCEGraphRemovals`, `--minify-whitespace`)
3. Post-`metadce` (`AJSDCE`, `--minify-whitespace`)
4. Inside `minify_wasm_imports_and_exports`
(`applyImportAndExportNameChanges`, `--minify-whitespace`)

Have `metadce` and `minify_wasm_imports_and_exports` return their
removal/renaming metadata instead of invoking `acorn_optimizer`
directly, and consolidate `applyDCEGraphRemovals`, post-`metadce`
`AJSDCE`, `applyImportAndExportNameChanges`, `stripDefaultUndefined`,
and `--minify-whitespace` into a single final `acorn_optimizer` pass.

Split out from #27926
@sbc100

sbc100 commented Oct 8, 2026

Copy link
Copy Markdown
Collaborator Author

Nice!

This looks right to me. Maybe the only question I have is if we want to run --compress in -O2 or just -O3 (I would lean towards not in -O2).

Really? Why? Do you expect -O2 output to be readable? Remember can always add -O2 --minify=0 if you prefer.

@sbc100

sbc100 commented Oct 8, 2026

Copy link
Copy Markdown
Collaborator Author

Also, we also run whitespace minification on main on -O2.. so changing that to not minify now seems odd.

@sbc100

sbc100 commented Oct 8, 2026

Copy link
Copy Markdown
Collaborator Author

Rebased on top of #27928 .. much smaller change now

@sbc100
Replace stripDefaultUndefined with Terser --compress
339d9e0 
Instead of a dedicated `stripDefaultUndefined` Acorn AST pass (added in
compression (`{defaults: false, evaluate: true, keep_fargs: false}`)
when minifying JS without Closure Compiler. In addition to stripping
`= undefined` default parameter values, this performs lightweight Terser
cleanups (`undefined` -> `void 0`, single-statement block brace
removal, and constant folding).

In general, this prevents us from needing to replicate compression
passes that already exist in upstream Terser (see emscripten-core#27871).

Timing highlights (`test/hello_world.c` with `-sINCLUDE_FULL_LIBRARY`,
~252 KB minified JS):
- `minify_sync` with `compress: false`: 105.5 ms
- `minify_sync` with `compress: {defaults: false, evaluate: true,
  keep_fargs: false}`: 188.3 ms (+82.8 ms, or ~+63 ms net after
  removing the `stripDefaultUndefined` AST walk), saving ~1.4 KB.
- On a standard `-O2` `hello_world.c` build (~8.5 KB JS), `minify_sync`
  goes from 3.2 ms to 5.5 ms (+2.3 ms).

This change, in general, should prevent us from being tempted to
replicate compression passes that already exist in upstream terser
(See emscripten-core#27871, for example).

See: emscripten-core#27839
@sbc100
sbc100 requested a review from kripken October 9, 2026 04:58
@sbc100

sbc100 commented Oct 9, 2026

Copy link
Copy Markdown
Collaborator Author

Should be good to go now.

@kripken

kripken commented Oct 9, 2026

Copy link
Copy Markdown
Member

Really? Why? Do you expect -O2 output to be readable? Remember can always add -O2 --minify=0 if you prefer.

I meant compile time, not readability. That is, I am usually ok with making -O3 slower when it helps output quality, but I think -O2 is a compromise that should be faster? In between -O1 and -O3

kripken
kripken approved these changes Oct 9, 2026
View reviewed changes

@kripken kripken left a comment

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

But this seems fast enough for -O2 as well.

@sbc100
sbc100 merged commit 30df0e8 into emscripten-core:main Oct 9, 2026
41 checks passed
@sbc100
sbc100 deleted the terser-compress branch October 9, 2026 18:18
@sbc100 sbc100 mentioned this pull request Oct 9, 2026
Open
sbc100 added a commit to sbc100/emscripten that referenced this pull request Oct 9, 2026
@sbc100
Use const in JS library code and enable Terser reduce_vars/unused
5fc14dc 
Enable `reduce_vars` and `unused` in Terser `--compress` (added in
emscripten-core#27926) in `tools/acorn-optimizer.mjs`.

This inlines `const` values into their expression uses and drops the
resulting unused local `const` declarations (along with unused function
parameters and single-use inner functions) in optimized builds that do
not run Closure Compiler.

This allows us to use `const` in JS library code without paying a code
size cost, while reducing total JS code size in
`test_no_closure_code_size` by 612 bytes.
sbc100 added a commit to sbc100/emscripten that referenced this pull request Oct 9, 2026
@sbc100
Use const in JS library code and enable Terser reduce_vars/unused
55b8b32 
Enable `reduce_vars` and `unused` in Terser `--compress` (added in
emscripten-core#27926) in `tools/acorn-optimizer.mjs`.

This inlines `const` values into their expression uses and drops the
resulting unused local `const` declarations (along with unused function
parameters and single-use inner functions) in optimized builds that do
not run Closure Compiler.

This allows us to use `const` in JS library code without paying a code
size cost, while reducing total JS code size in
`test_no_closure_code_size` by 612 bytes.

Timing highlights (`test/hello_world.c` with `-sINCLUDE_FULL_LIBRARY`,
~247 KB minified JS):
- `minify_sync` with `compress: {defaults: false, evaluate: true,
  keep_fargs: false}`: 165.1 ms (250,514 B)
- `minify_sync` adding `reduce_vars: true, unused: true`: 224.3 ms
  (+59.2 ms, or +26.7 ms for `unused` and +28.8 ms for `reduce_vars`),
  saving ~3.2 KB (247,346 B).
- On a standard `-O2` `hello_world.c` build (~8.4 KB JS), `minify_sync`
  goes from 3.9 ms to 6.2 ms (+2.3 ms), saving 96 B.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@kripken kripken kripken approved these changes
@dschuff dschuff Awaiting requested review from dschuff

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

2 participants

@sbc100 @kripken