jcs
/subtext
/amendments
Amendments on November 14, 2023
binkp: Abort idle connections, respect binkp_delete_done for sending
Keep track of the time we last received a frame from the server and
if nothing happens in a while, kill the connection.
Use the binkp_delete_done setting to move processed outbound packets
into the processed folder for debugging.
Also send the actual mtime of the packet we're sending instead of
1, since it can be helpful for the binkp server operator.
Amendments on November 15, 2023
binkpc: Bail when receiving an M_SKIP
It's supposed to be non-destructive and we'll retry on the next
connection, though because the file remains in the outbox, we'll
pick it up quite quickly and retry sooner than the poll interval.
jcs made amendment
543
5 months ago
jcs made amendment
544
5 months ago
Amendments on November 16, 2023
jcs made amendment
545
5 months ago
binkp: Remember when we failed connecting and delay next poll
If we had messages in the outbox and triggered a connection which
then failed, the main periodic loop will just pick up those outbox
packets right away and try reconnecting. In that case, don't look
for outbox packets and just use the next poll time.
jcs made amendment
547
5 months ago
jcs made amendment
548
5 months ago
fidopkt: Set proper product code and major version number
According to FTSC docs, we're supposed to set a product code (low)
of 0xfe if we haven't been assigned a number, which are now 16-bit.
Once that happens, low will be set to 0xff and high will become our
assigned number.
Amendments on November 17, 2023
jcs made amendment
550
5 months ago
jcs made amendment
551
5 months ago
jcs made amendment
552
5 months ago
Amendments on November 19, 2023
console: Don't panic when cursor is out of bounds, just clamp it
This is how our terminal size probing works, after all.
console: Revert previous, this was just an off-by-one
Clamping is done during CSI handling, so if we get here and we're
still out of bounds, this should panic.
Amendments on November 21, 2023
jcs made amendment
555
5 months ago
jcs made amendment
556
5 months ago
jcs made amendment
557
5 months ago
jcs made amendment
558
5 months ago
jcs made amendment
559
5 months ago
board: For new FTN messages posted locally, just toss them
Instead of writing the bile object ourself, use the toss function
which will do all of that along with the important bit of putting
the message id into the msgid cache. Previously we weren't doing
that so if the hub sent us back a copy of our own message, our
duplicate detection wasn't finding the existing msgid in our cache,
causing duplicate copies of all locally posted messages.
We still need to update this msgid cache when deleting FTN messages,
but we'll do that later.