| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
| |
Simplify primitive interface to read from a socket: now returns just a
single character.
|
|
|
|
|
| |
Now we can just read a single character from the socket, and it isn't
much slower either (maybe 10%).
|
|
|
|
| |
Make sure we clear the socket variable after closing it.
|
| |
|
|
|
|
|
|
|
| |
Fix CI. 3 different memory leaks in the socket internals.
The hard one was recognizing the need for
`receive-from-client-socket-and-close`.
|
| |
|
|
|
|
| |
Shorter branches above longer ones.
|
|
|
|
| |
This is just the same as tests for a fake file-system.
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This time we're opening real sockets so we might run into issues on
machines with firewalls. Macs for example flash up a dialog warning
people about a port being opened, though it shuts down immediately if
the test passes.
On the flip side, the test has greater verisimilitude. You don't really
need fakes except when you want to pin down the environment a test runs
in. The only way this test might be flaky is on a machine that has lots
of sockets open (so the random port opened by the test is in use for
something else). That and the firewall concern above. Hmm.
|
|
|
|
| |
Its interface now mirrors that of 'read-event'.
|
|
|
|
|
| |
Change the interface for reading a URL slightly so that we can directly
use the path in `assume-resources`.
|
|
|
|
| |
Fix CI. Revert accidentally-added files.
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
In the process I've also altered the API of $read-from-socket to return
a boolean (eof?) rather than the number of bytes read (which is implicit
in the length of the returned array).
|
|
|
|
| |
The bugfix of commit 3517 allows us to drop this redundant condition.
|
| |
|
| |
|
| |
|
| |
|
|
|
|
| |
Fix a few names in comments.
|
| |
|
| |
|
|
Testable network interface and write flow.
|