Stream-based sanitizer interface

[eschultink]

wip draft for sake of awareness.

provides little value except in NDJSON response cases, which to date are very few cases (slack analytics API, maybe glean audit log in future). in regular JSON cases, we're going to parse entire response into mem anyways.

Features

• work towards a fully Stream-based sanitizer interface; in theory, this will improve mem issues and avoid possible OOM in some cases.

Change implications

• dependencies added/changed? no
• something important to note in future release notes? no
• breaking changes? no

---

• To see the specific tasks where the Asana app for GitHub is being used, see below:
  • https://app.asana.com/0/0/1210711380911437

[eschultink]

so this check covers the Slack Analytics API case, but only that.

my confusion is how we're handling gzip in other cases. best guess is there must be some other layer at which we're decompressing. Either we do it explicitly somewhere else, or Google HTTP client is automatically seeing the Content-Encoding header and dealing with it. A concern is be that if indeed Google HTTP client is decompressing gzip but not stripping the Content-Encoding header, we're going to double-decompress the content