Commit 73cc00e reduced the number of emitted source locations a great
deal but it resulted in at least one observable regression:
export default async function f() {
return "abc" + x
}
f() // ReferenceError should point to 2:20 but pointed to 1:1
Emit source locations for expressions again. Increases the average
number of source locations by about 15%. Non-scientifically tested
by counting source locations emitted when parsing the test suite
before and after.
No test because we currently cannot easily test stack traces coming
from module imports.
Fixes: https://github.com/quickjs-ng/quickjs/issues/779
We were emitting gobs of inefficient bytecode that created an arguments
array on the stack, then applied it to the parent constructor.
Add a new opcode for initializing a derived class. Speeds up construction
by 500%, although sadly that is not visible in the web-tooling-benchmark,
only in micro-benchmarks.
Fixes: https://github.com/quickjs-ng/quickjs/issues/752
Emit source locations manually for more precise tracking. Don't infer
them from emitted bytecode opcodes because that leads to inaccurate
and sometimes surprising results.
Speeds up code generation (although infinitesimally) as a bonus.
Fixes: https://github.com/quickjs-ng/quickjs/issues/236
This commit implements resizable ArrayBuffers - RABs for short - and
extends typed arrays (TAs) to support fixed-length and length-tracking
modes.
SharedArrayBuffers (SABs) also support the maxByteLength option now but
I cheated and allocate all memory upfront because atomically resizing
memory allocations is hard and this commit is already big and complex.
The lion's share is updating all the TA prototype methods to deal with
RABs resizing underneath them. Method arguments can be arbitrary objects
with arbitrary .valueOf methods and arbitrary side effects, like...
resizing the RAB we're currently operating on.
Fixes: https://github.com/quickjs-ng/quickjs/issues/477
- `-s` strips the source code
- `-ss` strips source and line/column numbers information
- `qjsc repl.js` generates an object size of **105726** bytes
- `qjsc -s repl.js` generates an object size of **20853** bytes
- `qjsc -ss repl.js` generates an object size of only **16147** bytes
- compile repl.js with `-ss`
- bump byte code version to 12
- improve `JS_DumpString`: use `L` prefix for wide strings
- dump variable kind and flags for locals and closures
- disassemble byte code in DUMP_READ_OBJECT
- pass start_pos to `dump_byte_code` and `dump_single_byte_code`
- write constant pool before function bytecode (bump version to 11)
- update generated code
Translate IC opcodes to their non-IC variants before writing them out.
Before this commit they were not byte-swapped properly, breaking the
ability to load serialized bytecode containing ICs on systems with
different endianness. Inline caches are recomputed as needed now.
A pleasing side effect of this change is that serialized bytecode is,
on average, a little smaller because fewer atoms are duplicated now.
