James 'zofrex' Sanderson

@mbj this should be an easy one to review and merge :)

@mbj oops, my bad! Sent that to you now, sorry, was just going through old open PRs and forgot about the CLA for this one!

I can no longer reproduce the crash in the second example. Of course, the contents of the produced IcmpPacket are nonsensical but that's to be expected from the primary issue...

As a library implementer I'm curious - what's non-standard about the second code? I see that the ordering of the set/faction lists within in the two copies list is different,...

@Billzabob I don't see an ordering specified for equal-length set/faction lists though? > The set/faction lists are ordered by increasing length They are, here - the ones in different orders...

Based on [the PoC](https://github.com/rubygems/rubygems/pull/5911) the smoothest way to add this was as a new section in the lockfile, which would look something like this: ``` EDIT WARNING Do not edit...

Thank you for your comments! > the section feels artificial Agreed! It doesn't actually _have_ to match the other sections in style, it could look like this for example: ```...

> In bundle outdated we have a table output, maybe we could use some inspiration from there. Good idea, I think it makes a lot of sense for these two...

As a temporary workaround you can fix this in any project using a Gemfile by limiting the version of haml in your Gemfile, for example: ``` source "https://rubygems.org" gem 'middleman',...

I think the url crate may be the issue here, not the script, but I'm less than certain. The [url crate's Cargo.toml](https://github.com/servo/rust-url/blob/f99996a6f2196741d6f9d7f33f8630d7172af316/Cargo.toml): ``` [[test]] name = "format" [[test]] name =...