3 Jul 2024 |
| rvbcldud joined the room. | 21:36:41 |
rvbcldud | Redacted or Malformed Event | 21:43:46 |
8 Jul 2024 |
smoltcp-bot | New PR: Use assert!() to check prefix_len in Ipv4Cidr::new | 22:39:51 |
9 Jul 2024 |
smoltcp-bot | PR merged: Use assert!() to check prefix_len in Ipv4Cidr::new | 07:13:01 |
12 Jul 2024 |
smoltcp-bot | New PR: Fix benchmark and update results | 18:48:36 |
13 Jul 2024 |
smoltcp-bot | PR merged: Don't delay ACKs for significant window updates | 05:03:16 |
smoltcp-bot | PR merged: Fix benchmark and update results | 05:04:04 |
14 Jul 2024 |
smoltcp-bot | New PR: Reset retransmit timer on ack regardless of the current timer state | 19:07:41 |
15 Jul 2024 |
smoltcp-bot | New PR: Limiting the number of poll loops to prevent DoS events | 09:30:11 |
18 Jul 2024 |
smoltcp-bot | PR closed without merging: RFC: support for TCP Segmentation Offloading (TSO) | 10:58:58 |
dirbaio | took 1 year to think about it | 11:14:40 |
smoltcp-bot | New PR: [Breaking] Update to bitflags 2.x | 13:06:50 |
22 Jul 2024 |
dirbaio | is this something we should fix in smoltcp instead of embassy? | 20:15:22 |
dirbaio | https://github.com/embassy-rs/embassy/pull/3198 | 20:15:23 |
dirbaio | when a socket is reset, data in the tx queue stays there foreve.r | 20:15:49 |
dirbaio | * when a socket is reset, data in the tx queue stays there forever. | 20:15:50 |
dirbaio | it feels odd to keep data that will never get sent, perhaps smoltcp should clear the tx data on reset | 20:16:31 |
dirbaio | * it feels odd to keep data that will never get sent, perhaps smoltcp should clear the tx queue on reset | 20:16:34 |
dirbaio | otoh this is not an issue until you try to use "is tx queue empty" to build some kind of "flush" like embassy does... so maybe it is embassy's "fault" | 20:17:08 |
Robert Jördens | In #351 you enabled distinguishing "received all the data" from "received data may be truncated". Seems natural to support a similar distinction for TX. I wonder if this is the only scenario where tx data can still be in the queue when reaching Closed . | 22:29:20 |
23 Jul 2024 |
smoltcp-bot | PR merged: Limiting the number of poll loops to prevent DoS events | 14:13:31 |
Catherine | In reply to @dirbaio:matrix.org it feels odd to keep data that will never get sent, perhaps smoltcp should clear the tx queue on reset this sounds very reasonable | 19:27:35 |
Catherine | there are some protocols which you could conceivably want to intentionally work that way, but I don't think we should support that use case | 19:27:51 |
dirbaio | which protocols? | 19:28:08 |
dirbaio | for rx sure, you might still want to process the data you did receive after a reset, but for tx.? | 19:28:22 |
Catherine | something self-synchronizing | 19:31:21 |
Catherine | SCPI works kind of like this in many instruments, horrifically | 19:31:38 |
Catherine | it shouldn't, but it does. (I'm stating the existence of this use pattern, not arguing for it--against it in fact) | 19:32:00 |
24 Jul 2024 |
smoltcp-bot | New PR: Fix performance related pedantic clippy lints | 18:51:18 |
25 Jul 2024 |
smoltcp-bot | PR closed without merging: Fix performance related pedantic clippy lints | 21:22:59 |