11:12 <whitequark> sb0: I believe I've nailed down the correct approach for fixing #1007

11:12 <GitHub41> [smoltcp] whitequark commented on issue #259: Yep, I will handle this soon-ish. https://github.com/m-labs/smoltcp/issues/259#issuecomment-423736743

11:15 <GitHub181> [smoltcp] whitequark commented on issue #261: This is interesting. The logic in `Assembler::add` is not trivial. Do you think you could record a trace under `rr` so that I can step back through it? https://github.com/m-labs/smoltcp/issues/261#issuecomment-423736883

11:16 <GitHub60> [smoltcp] whitequark commented on issue #51: @podhrmic This is done, right? https://github.com/m-labs/smoltcp/issues/51#issuecomment-423736978

11:22 <GitHub-m-labs> [artiq] jordens commented on issue #1155: Yes. It doesn't help. https://github.com/m-labs/artiq/issues/1155#issuecomment-423737345

11:25 <GitHub85> [smoltcp] pothos commented on issue #261: That's a good idea - or I try to reproduce this in a test case. What do you think would happen in a release compile? https://github.com/m-labs/smoltcp/issues/261#issuecomment-423737471

11:28 <GitHub171> [smoltcp] whitequark commented on issue #261: I have no idea, but probably data loss--this assert is there for a good reason. We actually might want to elevate it to `assert!`. https://github.com/m-labs/smoltcp/issues/261#issuecomment-423737658

11:35 <GitHub63> [smoltcp] pothos commented on issue #261: The while loop can be left if `index == self.contigs.len()`. What should be the result in this case? https://github.com/m-labs/smoltcp/issues/261#issuecomment-423738005

11:47 <GitHub91> [smoltcp] whitequark commented on issue #261: I don't actually know! I wrote this code a while ago and I don't remember the exact rationale for that condition. How about adding a fuzzer-based test that generates and feeds the assembler random data ranges and sees what happens? That should discover a reduced test case for the crash very rapidly. https://github.com/m-labs/smoltcp/issues/261#issuecomment-423738655

11:58 <GitHub190> [smoltcp] pothos commented on issue #261: I'm new to reading it as well and reported this here in case others also observe it. By reading quickly, I'm very suspicious about this part because it can explain why suddenlty the 1-length packet was not dropped anymore:... https://github.com/m-labs/smoltcp/issues/261#issuecomment-423739161

11:59 <GitHub72> [smoltcp] pothos commented on issue #261: I'm new to reading it as well and reported this here in case others also observe it. By reading quickly, I'm very suspicious about this part because it can explain why suddenly the 1-length packet was not dropped anymore:... https://github.com/m-labs/smoltcp/issues/261#issuecomment-423739161

12:01 <GitHub96> [smoltcp] whitequark commented on issue #261: Excellent catch! Can you write a testcase? https://github.com/m-labs/smoltcp/issues/261#issuecomment-423739329

12:03 <GitHub166> [smoltcp] pothos commented on issue #261: I'll have to dive into it a bit more, then yes ;) https://github.com/m-labs/smoltcp/issues/261#issuecomment-423739434

13:01 <GitHub128> [smoltcp] pothos opened pull request #264: Increase number of assembler gaps to 32 (master...assembler-holes) https://github.com/m-labs/smoltcp/pull/264

13:22 <GitHub53> [smoltcp] whitequark commented on issue #264: 64 contigs means 128 machine words per TCP socket, so usually 0.5K. This is quite a bit for the smallest targets. How about *either* a preallocated set of 4 gaps (for the smallest targets) *or* a heap-allocated assembler buffer (if you can spare 0.5K that'll usually go unused you probably have `Box` already)? https://github.com/m-labs/smoltcp/pull/264#issuecomment-4237

13:56 <GitHub154> [smoltcp] pothos opened pull request #265: Only change the assembler state on success (master...assember_rejected_add_keeps_state) https://github.com/m-labs/smoltcp/pull/265

14:00 <GitHub59> [smoltcp] whitequark commented on issue #265: @m-labs-homu r+... https://github.com/m-labs/smoltcp/pull/265#issuecomment-423746166

14:00 <GitHub87> [smoltcp] m-labs-homu commented on issue #265: :pushpin: Commit 9ccf18e has been approved by `whitequark`

14:00 <GitHub74> [smoltcp] m-labs-homu pushed 1 new commit to auto: https://github.com/m-labs/smoltcp/commit/68869defc79e7bce8d4a7e09aa7981ccb7cc4a44

14:00 <GitHub74> smoltcp/auto 68869de Kai Lüke: Only change the assembler state on success...

14:00 <GitHub167> [smoltcp] m-labs-homu commented on issue #265: :hourglass: Testing commit 9ccf18e0b61efc49480aa538495a7e746bed1d16 with merge 68869defc79e7bce8d4a7e09aa7981ccb7cc4a44... https://github.com/m-labs/smoltcp/pull/265#issuecomment-423746207

14:12 <GitHub164> [smoltcp] m-labs-homu merged auto into master: https://github.com/m-labs/smoltcp/compare/60db9424e039...68869defc79e

14:12 <GitHub3> [smoltcp] m-labs-homu closed pull request #265: Only change the assembler state on success (master...assember_rejected_add_keeps_state) https://github.com/m-labs/smoltcp/pull/265

14:12 <travis-ci> m-labs/smoltcp#1170 (auto - 68869de : Kai Lüke): The build passed.

14:12 <GitHub56> [smoltcp] m-labs-homu commented on issue #265: :sunny: Test successful - [status-travis](https://travis-ci.org/m-labs/smoltcp/builds/431871967?utm_source=github_status&utm_medium=notification)

14:12 <travis-ci> Change view : https://github.com/m-labs/smoltcp/compare/60db9424e039...68869defc79e

14:13 <travis-ci> Build details : https://travis-ci.org/m-labs/smoltcp/builds/431871967

14:23 <travis-ci> m-labs/smoltcp#1171 (master - 68869de : Kai Lüke): The build passed.

14:23 <travis-ci> Change view : https://github.com/m-labs/smoltcp/compare/60db9424e039...68869defc79e

14:23 <travis-ci> Build details : https://travis-ci.org/m-labs/smoltcp/builds/431874491

14:46 <GitHub19> [smoltcp] pothos commented on issue #264: Rather an API change for Assembler and ManagedSlice or really just a Box for std/alloc? https://github.com/m-labs/smoltcp/pull/264#issuecomment-423749074

14:47 <GitHub117> [smoltcp] whitequark closed issue #261: Debug panic in packet assembler https://github.com/m-labs/smoltcp/issues/261

14:52 <GitHub198> [smoltcp] pothos commented on issue #264: Rather an API change for Assembler and use ManagedSlice or really just use a Box for std/alloc? https://github.com/m-labs/smoltcp/pull/264#issuecomment-423749074

18:22 <GitHub175> [smoltcp] jhwgh1968 commented on issue #256: A quick update: I have made some progress on this, but am running into blocking issues when creating unit tests.... https://github.com/m-labs/smoltcp/issues/256#issuecomment-423763530

18:58 <GitHub162> [smoltcp] pothos commented on issue #264: Seems that `cargo test` has a bug in evaluating the cfg macros: There is a compile error because some definitions are compiled with cfg std and some without!? https://github.com/m-labs/smoltcp/pull/264#issuecomment-423765754

20:19 <GitHub1> [smoltcp] whitequark commented on issue #264: I think `#[test]` implies `std`. https://github.com/m-labs/smoltcp/pull/264#issuecomment-423770542