<bertrand.lupart[at]linkeo.com> (Bertrand LUPART) wrote:
> mixed result = catch{ http=Protocols.HTTP.get_url(url); };
That is a synchronous query. The timeouts only apply to the
asynchronous ones (done through e.g. Protocols.HTTP.do_async_method).
I agree that it's confusing that the synchronous variety doesn't obey
the timeouts. It probably has historic reasons since it was very hard
to do blocking I/O with timeouts in pike before. Nowadays it's still a
bit tricky, but it's possible to emulate blocking I/O with nonblocking
and a local backend. The synchronous requests ought to be fixed to do
it that way.
|