Yikes - that far exceeds my abilities. I don't know if anyone else can
recreated this issue but it sounds like a job for someone with time on
their hands, the necessary skills & knowledge, and likes a difficult
puzzle. When I started this thread I was hoping for "change this
line...in this file" or something equally manageable for myself :-)
On 2022-09-20 07:53, Crystal Kolipe wrote:
So the next step would be to either:
1. Read through all of the javascript, work out which requests should be
made, and compare that list to the actual list of requests made.
or
2. Compare the requests made on a setup which is known to work, with
the requests made on a setup which does not work.
Presumably either something is not being sent from the client, or
something is not received, (or is received too late, after a timeout has
expired).
Or maybe it is being sent as far as the client is concerned, but then
gets mangled or delayed at the IP layer.
Getting that far is complicated by the fact that there could be more than
one issue, more than one request not being serviced, that the requests
might be answered by different machines at the remote end, and there might
also be some intentional randomisation going on between otherwise
'identical' login attempts.
If you do manage to work out which javascript-initiated communication is
not going through, you then need to find the _real_ reason why.
It could be something seemingly completely unrelated. Maybe deep within
that javascript is a function that is also implemented in web-assembly
and that version is being used on Firefox but not on Surf, (which doesn't
support it), causing a slight timing difference.
The list of possibilities is almost endless at this point.
