Name: massive-heap-snapshot-parser
Owner: NodeSource
Description: small library for parsing massively massive snapshots and giving you a way to access its contents sanely
Created: 2017-01-12 18:31:18.0
Updated: 2017-12-29 11:39:34.0
Pushed: 2017-01-13 00:35:36.0
Homepage: null
Size: 21
Language: JavaScript
GitHub Committers
User | Most Recent Commit | # Commits |
---|
Other Committers
User | Most Recent Commit | # Commits |
---|
To install just do:
m install mhsp
This API will extract "snapshot"
, "nodes"
, "edges"
and "strings"
from
very large snapshots in a way that allows you to continue working with them. It
doesn't however show "trace_function_infos"
, "trace_tree"
or "samples"
.
Mainly because I don't need them right now. File an issue or open a PR if you'd
like to help.
parseSnapshot(path)
DataAccessor
instanceThis does all the magic. Just give it a path to a snapshot and it'll automatically do everything for you. Here's an example:
strict';
t { parseSnapshot } = require('mhsp');
t accessor = parseSnapshot('./my-big.heapsnapshot');
Warning that for large snapshots this can take several minutes.
accessor.snapshot
Object
This is the metadata object that contains various information about the snapshot.
accessor.nodes
Uint32Array
This is a Uint32Array
of all the nodes in the snapshot. Go ahead and access
it by index.
accessor.edges
Uint32Array
This is also a Uint32Array
that can be accessed by index.
accessor.getString(index)
String
Returning strings from the "strings"
field isn't as straight forward. Some
fun index tracking is done under the hood so the entire "strings"
section can
live in one big Buffer
.
accessor.writeToFile(path)
Number
of bytes written to diskSince generating the accessor takes so long you can go ahead and write the
entire thing to disk in the form of a binary blob. Can then use importBin()
to read it back in later. Is much much faster.
importBin(path)
DataAccessor
instanceRead in files that have already been processed and written to disk as binary blobs. On my i7 the first time I process a 2GB snapshot can take over 1.5 minutes. But reading it back in this way only takes a few seconds. Highly recommended.