EMSCRIPTEN_FETCH_REPLACE needed for loads to memory?

[kripken]

The testcase here:

https://github.com/emscripten-core/emscripten/compare/fetch-wasm-pthread-async?expand=1

passes as is, but if EMSCRIPTEN_FETCH_REPLACE is removed then it fails. That seems odd since the docs say

// If the file already exists in IndexedDB, the old file will be deleted and a new download is started.

But the test does not do anything with IndexedDB, it just fetches to memory. Are the docs out of date, or does REPLACE mean something here?

cc @juj

[stale]

This issue has been automatically marked as stale because there has been no activity in the past year. It will be closed automatically if no further activity occurs in the next 7 days. Feel free to re-open at any time if this issue is still relevant.

[VirtualTim]

I believe this is still relevant.

[stale]

This issue has been automatically marked as stale because there has been no activity in the past year. It will be closed automatically if no further activity occurs in the next 30 days. Feel free to re-open at any time if this issue is still relevant.

[heshuimu]

This is still relevant even now. The workaround in #8749 still works which is not expected.

[sbc100]

From my reading of the docs EMSCRIPTEN_FETCH_LOAD_TO_MEMORY does mean "don't use indexDB" .. it means that the full result will be available in memory after the request. The request can still be satisfied by an indexDB read. Is that right @juj?

Still it seems like if EMSCRIPTEN_FETCH_SYNCHRONOUS is not compatible with indexDB at all then we shouldn't trying to read or write to it in this mode.