I don't have much experience with TaskRunners or Sequences, but I'm being asked to migrate stuff. One question I keep running into: what's the difference between "thread-safe" and "sequenced"? And how is something that is "not thread-safe", like base::WeakPtr, okay to use in a SequencedTaskRunner (which doesn't even guarantee thread-safety)?

Specifically, these two statements in the docs don't sound compatible. [1]

"Sequences inherently provide thread-safety."
"sequences can hop threads instead of being stuck behind unrelated work on a dedicated thread"

If a sequence runs tasks on different threads, how is it thread-safe? It does seem that, given some very complicated conditions[2], a SequencedTaskRunner "provides most (but not all) of the same guarantees" as "tasks are run on a single dedicated thread".

In the example given for base::WeakPtr [3], does the example's correctness rely on the conditions expressed in SequencedTaskRunner's docs in [2]? That is, do Compute and Store need to be nestable/non-nestable, be guaranteed not to post other tasks, and so forth? It seems like it would be a lot of work to ensure that each function, and the functions it calls, etc. never does any of those things.

Should the docs be expanded to explain these concepts, or do we need a multithreading-noob-friendly intro version?

[1] https://chromium.googlesource.com/chromium/src/+/master/docs/threading_and_tasks.md
[2] https://cs.chromium.org/chromium/src/base/sequenced_task_runner.h?type=cs&sq=package:chromium&q=sequencedtaskrunner&l=38-65
[3] https://chromium.googlesource.com/chromium/src/+/master/docs/threading_and_tasks.md#Using-base_WeakPtr