@Sophia-Gold, thanks for your PR! By analyzing the history of the files in this pull request, we identified @7c6f434c, @bjornfor and @Fuuzetsu to be potential reviewers.

Per CONTRIBUTING.md commit messages should be formatted:
(pkg-name | service-name): (from -> to | init at version | refactor | etc)
In this case it should be:
simgrid: 3.11.1 -> 3.15

I think you changed the PR title instead of the commit message

Also, in general such updates should be submitted against master first

Apologies. I neglected to create a new branch after checking out the master and therefore decided to make this commit through the GUI, which explains some of the oddness. I'm not sure how to amend the message now, if that's even preferable.

Cherry-picked into master and release-17.03

Thank you.

Command-line git commit --amend does work from this state, I guess it would be possible to pull the PR branch, amend it and force-push. We have a complicated workflow in terms of initially recommending using channels for installation and branch tops for contributions.

Thanks for the update.

For reference, sandboxed builds got broken by this, but it should be OK now.

That's odd. I built it on NixOS with a local path and got none of these errors. Glad to hear they're resolved, though.

Well, look at the fix — you probably have a global /usr/bin/env which can be accessed unless sandboxing is used. Well, you can enable sandboxing by default on your system if you prefer (purer/more reproducible builds, somewhat longer initialisation of a build with many dependencies)

Got it. Next time I'll enable sandboxing when testing a new package.

Most contributors use sandboxing all the time, I think. There's just a split-second penalty per derivation, so that's totally worth it for me.