<whitequark>
bb-m-labs: force build --props=package=rustc conda-lin64
<bb-m-labs>
build forced [ETA 37 seconds]
<bb-m-labs>
I'll give a shout when the build finishes
<GitHub95>
[smoltcp] whitequark commented on issue #44: Quick comment so this doesn't sit idle: can you please add a `struct Ipv4Cidr { addr, mask }` and `enum IpCidr { Ipv4(Ipv4Cidr) }`, just like it's done for `struct Ipv4Addr` and `enum IpAddr`? I'd really like smoltcp to gain these abstractions eventually, anyway. https://git.io/vdTB5
<GitHub42>
[smoltcp] whitequark commented on issue #44: Oh sorry I already wrote this comment and forgot about it! Let me take another look. https://git.io/vdTBb
<GitHub53>
[smoltcp] whitequark commented on issue #44: Oh sorry I already wrote this comment and forgot about it! If you don't want to deal with this I can implement it myself, no big deal. https://git.io/vdTBb
<GitHub144>
[smoltcp] batonius commented on issue #44: My first thought was to implement some kind of `Cidr` data type, which is handy to have anyway, but I decided against it because it doesn't fit here conceptually. Specifically, 'mask length' is an attribute of an address assigned to an interface, not of the default route, which is a bare IP address, so it would be wrong to bundle them together in a single `Cidr` value. That means the right approach shoul
<GitHub44>
[smoltcp] whitequark commented on issue #44: > That means the right approach should probably involve changing EthernetInterface::new to accept a slice of IpCidrs instead of IpAdddresss and require an IpAddress of the default gateway as well, but I really didn't want to change public API.... https://git.io/vdT27
<GitHub148>
artiq/master 7799413 whitequark: firmware: update Rust from 1.18.0 to 1.20.0.
<GitHub46>
[smoltcp] whitequark commented on pull request #44 adcf30f: Nit: I think `MT` used to mean something but now it can be just replaced with `T` everywhere here (so `GatewayT`). https://git.io/vdTiD
<GitHub106>
[smoltcp] whitequark commented on pull request #44 adcf30f: I assume there's an `impl Into<Option<T>> for Option<Into<T>>` or something like it? https://git.io/vdTiM
<GitHub87>
[smoltcp] whitequark commented on pull request #44 adcf30f: Prefix length, not mask length; the mask is as long as the IP address, but has only as many zeroes as the prefix length. https://git.io/vdTiS
<GitHub91>
[smoltcp] whitequark commented on pull request #44 adcf30f: Nit: just `contains`. https://git.io/vdTiy
<GitHub69>
[smoltcp] whitequark commented on pull request #44 adcf30f: Nit: I think `address` sounds nicer, and see above on mask_len. https://git.io/vdTiH
<GitHub40>
[smoltcp] whitequark commented on pull request #44 adcf30f: This function doesn't seem very useful as the valid ranges of prefix lengths depend on the address. https://git.io/vdTi9
<GitHub176>
[smoltcp] whitequark commented on pull request #44 adcf30f: Since this field is `pub`, the struct needs an `is_valid()` method, which should be checked in `set_protocol_addrs`. https://git.io/vdTid
<GitHub1>
[smoltcp] whitequark commented on pull request #44 adcf30f: I've been uncomfortable calling this data structure "CIDR" for a while but I just realized what the proper name for it would be: `IpSubnet`! That makes much more sense. I'm sorry for having you change everything. https://git.io/vdTiQ
<GitHub36>
[smoltcp] whitequark commented on pull request #44 adcf30f: Nit: the phrase "default gateway" is a remnant of the times where a single LAN segment would conceivably have more than one gateway. At this point it's an anachronism (like classful networks!), let's call this just "gateway". Also applies elsewhere in this PR. https://git.io/vdTi5
<GitHub172>
[smoltcp] whitequark commented on pull request #44 adcf30f: Nit: just `address`. https://git.io/vdTi7