@sc3d/stacktracey
v2.1.9
Published
Friendly fork of stacktracey to fix bugs.
Downloads
23
Maintainers
Keywords
Readme
StackTracey
Parses call stacks. Reads sources. Clean & filtered output. Sourcemaps. Node & browsers.
Why
- [x] Simple
- [x] Works in Node and browsers, *nix and Windows
- [x] Allows hiding library calls / ad-hoc exclusion (via
// @hide
marker) - [x] Provides source text for call locations
- [x] Fetches sources (via get-source)
- [x] Supports both asynchronous and synchronous interfaces (works even in browsers)
- [x] Full sourcemap support
- [x] Extracts useful information from
SyntaxError
instances - [x] Pretty printing
What For
- Error overlay UIs for easier front-end development
- Better error reporting for Node projects
- Advanced logging (displaying call locations)
- Assertion printing
How To
npm install stacktracey
import StackTracey from 'stacktracey'
Captures the current call stack:
stack = new StackTracey () // captures the current call stack
Parses stacks from an Error
object:
stack = new StackTracey (error)
stack = new StackTracey (error.stack) // ...or from raw string
Stores parsed data in .items
:
stack.items.length // num entries
stack.items[0] // top
...where each item exposes:
{
beforeParse: <original text>,
callee: <function name>,
calleeShort: <shortened function name>,
file: <full path to file>, // e.g. /Users/john/my_project/node_modules/foobar/main.js
fileRelative: <relative path to file>, // e.g. node_modules/foobar/main.js
fileShort: <short path to file>, // e.g. foobar/main.js
fileName: <file name>, // e.g. main.js
line: <line number>, // starts from 1
column: <column number>, // starts from 1
index: /* true if occured in HTML file at index page */,
native: /* true if occured in native browser code */,
thirdParty: /* true if occured in library code */,
hide: /* true if marked as hidden by "// @hide" tag */,
syntaxError: /* true if generated from a SyntaxError instance */
}
Accessing sources (synchronously, use with caution in browsers):
stack = stack.withSources () // returns a copy of stack with all items supplied with sources
top = stack.items[0] // top item
Accessing sources (asynchronously, preferred method in browsers):
stack = await stack.withSourcesAsync () // returns a copy of stack with all items supplied with sources
top = stack.items[0] // top item
...or:
top = stack.withSourceAt (0) // supplies source for an individiual item (by index)
top = await stack.withSourceAsyncAt (0) // supplies source for an individiual item (by index)
...or:
top = stack.withSource (stack.items[0]) // supplies source for an individiual item
top = await stack.withSourceAsync (stack.items[0]) // supplies source for an individiual item
The returned items contain the following additional fields (already mapped through sourcemaps):
{
... // all the previously described fields
line: <original line number>,
column: <original column number>,
sourceFile: <original source file object>,
sourceLine: <original source line text>
}
To learn about the sourceFile
object, read the get-source docs.
Cleaning Output
Synchronously (use with caution in browsers):
stack = stack.clean ()
...or (asynchronously):
stack = await stack.cleanAsync ()
It does the following:
- Reads sources (if available)
- Excludes locations marked with the
isThirdParty
flag (library calls) - Excludes locations marked with a
// @hide
comment (user defined exclusion) - Merges repeated lines (via the
.mergeRepeatedLines
)
You can customize its behavior by overriding the isClean (entry, index)
predicate.
Custom isThirdParty
Predicate
You can override the isThirdParty
behavior by subclassing StackTracey
:
class MyStackTracey extends StackTracey {
isThirdParty (path, externalDomain) { // you can use externalDomain to include traces from libs from other domains
return (super.isThirdParty (path) // include default behavior
|| path.includes ('my-lib')) // paths including 'my-lib' will be marked as thirdParty
&& !path.includes ('jquery') // jquery paths won't be marked as thirdParty
}
}
...
const stack = new MyStackTracey (error).withSources ()
Pretty Printing
const prettyPrintedString = new StackTracey (error).withSources ().asTable ()
const prettyPrintedString = (await new StackTracey (error).withSourcesAsync ()).asTable () // asynchronous version
...or (for pretty printing cleaned output):
const prettyPrintedString = new StackTracey (error).clean ().asTable ()
const prettyPrintedString = (await new StackTracey (error).cleanAsync ()).asTable () // asynchronous version
It produces a nice compact table layout (thanks to as-table
), supplied with source lines (if available):
at shouldBeVisibleInStackTrace test.js:25 const shouldBeVisibleInStackTrace = () => new StackTracey ()
at it test.js:100 const stack = shouldBeVisibleInStackTrace ()
at callFn mocha/lib/runnable.js:326 var result = fn.call(ctx);
at run mocha/lib/runnable.js:319 callFn(this.fn);
at runTest mocha/lib/runner.js:422 test.run(fn);
at mocha/lib/runner.js:528 self.runTest(function(err) {
at next mocha/lib/runner.js:342 return fn();
at mocha/lib/runner.js:352 next(suites.pop());
at next mocha/lib/runner.js:284 return fn();
at <anonymous> mocha/lib/runner.js:320 next(0);
If you find your pretty printed tables undesirably trimmed (or maybe too long to fit in the line), you can provide custom column widths when calling asTable
(...or, alternatively, by overriding maxColumnWidths ()
method):
stack.asTable ({
callee: 30,
file: 60,
sourceLine: 80
})
Using As A Custom Exception Printer In Node
You can even replace the default NodeJS exception printer with this! This is how you can do it:
process.on ('uncaughtException', e => { /* print the stack here */ })
process.on ('unhandledRejection', e => { /* print the stack here */ })
But the most simple way to achieve that is to use the ololog
library (that is built upon StackTracey and several other handy libraries coded by me). Check it out, it's pretty awesome and will blow your brains out :)
const log = require ('ololog').handleNodeErrors ()
// you can also print Errors by simply passing them to the log() function
Parsing SyntaxError
instances
For example, when trying to require
a file named test_files/syntax_error.js
:
// next line contains a syntax error (not a valid JavaScript)
foo->bar ()
...the pretty printed call stack for the error thrown would be something like:
at (syntax error) test_files/syntax_error.js:2 foo->bar ()
at it test.js:184 try { require ('./test_files/syntax_error.js') }
at runCallback timers.js:781
at tryOnImmediate timers.js:743
at processImmediate [as _immediat timers.js:714
...where the first line is generated from parsing the raw output from the util.inspect
call in Node. Unfortunately, this won't work in older versions of Node (v4 and below) as these versions can't provide any meaningful information for a SyntaxError
instance.
Array Methods
All StackTracey instances expose map
, filter
, concat
and slice
methods. These methods will return mapped, filtered, joined, reversed and sliced StackTracey
instances, respectively:
s = new StackTracey ().slice (1).filter (x => !x.thirdParty) // current stack shifted by 1 and cleaned from library calls
s instanceof StackTracey // true
Extra Stuff
You can compare two locations via this predicate (tests file
, line
and column
for equality):
StackTracey.locationsEqual (a, b)
To force-reload the sources, you can invalidate the global source cache:
StackTracey.resetCache ()
Projects That Use StackTracey
- Ololog — a better
console.log
for the log-driven debugging junkies! - CCXT — a cryptocurrency trading library that supports 130+ exchanges
- pnpm — a fast, disk space efficient package manager (faster than npm and Yarn!)
- panic-overlay — a lightweight standalone alternative to
react-error-overlay