The XCLAIM docs state the XCLAIM increments the delivery counter for
messages. This PR makes the code match the documentation - which seems
like the desired behaviour - whilst still allowing RETRYCOUNT to be
specified manually.
My understanding of the way streamPropagateXCLAIM() works is that this
change will safely propagate to replicas since retry count is pulled
directly from the streamNACK struct.
Fixes#5194
This commit fixes#5570. It is a similar bug to one fixed a few weeks
ago and is due to the range API to be called with NULL as "end ID"
parameter instead of repeating again the start ID, to be sure that we
selectively issue the entry with a given ID, or we get zero returned
(and we know we should emit a NULL reply).
This fixes the issue reported in #5570.
This was fixed the hard way, that is, propagating more information to
the lower level API about this being a request to read just the history,
so that the code is simpler and less likely to regress.
This bug had a double effect:
1. Sometimes entries may not be emitted, producing broken protocol where
the array length was greater than the emitted entires, blocking the
client waiting for more data.
2. Some other time the right entry was claimed, but a wrong entry was
returned to the client.
This fix should correct both the instances.
They play better with Lua scripting, otherwise Lua will see status
replies as "ok" = "string" which is very odd, and actually as @oranagra
reasoned in issue #5456 in the rest of the Redis code base there was no
such concern as saving a few bytes when the protocol is emitted.
Slaves and rebooting redis may have different radix tree struct,
by different stream* config options. So propagating approximated
MAXLEN to AOF/slaves may lead to date inconsistency.
If we rewrite the MAXLEN argument as zero when no trimming
was performed, date between master and slave and aof will
be inconsistent, because `xtrim maxlen 0` means delete all
entries in stream.
The conclusion, that a xread request can be answered syncronously in
case that the stream's last_id is larger than the passed last-received-id
parameter, assumes, that there must be entries present, which could be
returned immediately.
This assumption fails for empty streams that actually contained some
entries which got removed by xdel, ... .
As result, the client is answered synchronously with an empty result,
instead of blocking for new entries to arrive.
An additional check for a non-empty stream is required.
