1148163 - Clean up cc-analyzer.js

Status: RESOLVED WONTFIX

(Testing :: Mochitest, defect)

Description

[Andrew McCreight [:mccr8]]

Right now, this creates a JS version of the entire CC graph, but most of that information is not needed.  All the CC leak analyzer actually needs is a list of nsGlobalWindows that are not garbage, and their description, so it can extract the window id.  This cleanup isn't strictly needed, but it makes dealing with the file nicer.

Comment 1

[Andrew McCreight [:mccr8]]

Attachment: part 1 - Remove unused fields from cc-analyzer.

The CC analyzer looks for non-garbage nsGlobalWindows in the CC graph.
To do this, it needs to store a list of CC objects (along with their
description and whether they are garbage or not). Right now, the CC
analyzer is very general and stores its own copy of the entire graph,
but this generality hasn't been used in the last few years, so I think
it is reasonable to remove that to make things simpler. processingCount
isn't used at all.  That's what part 1 does.

Right now, the CC stores every CC thing it sees, then looks through
them all after the fact for ones matching a particular string.  Part 2
changes this so we store the string we're interested in, and only adds
things that match.  This should greatly reduce how much we store.

try run: https://treeherder.mozilla.org/#/jobs?repo=try&revision=8006219b661c

Comment 2

[Andrew McCreight [:mccr8]]

Attachment: part 2 - Only add matching nodes to the cc-analyzer.js graph.

Comment 3

[Andrew McCreight [:mccr8]]

I'm just going to remove the CCAnalyzer.