Support multiple backends
hkupty opened this issue · 0 comments
hkupty commented
Curl does the job but:
- involves communicating through stdout, therefore:
- it has only one channel to transmit the information
- To get HTTP statuscode, we'll have to parse the output
- It's a synchronous
system
call:- Can be made asynchronous but I guess that'll make things a bit more complicated;
- using
jobstart
might be an option, but then maybe it'll need to concat the output before parsing.
Supporting alternative backends will allow for:
- gradual migration;
- testing different approaches;
- ensuring a stable backend exists as alternative;
- falling back to curl if nothing is available;
- supporting other platforms (i.e. Windows)