Create a continuous integration server for building rustpkg packages

[brson]

I would like to know when packages break. We can possibly extend rustbot to do this.

[graydon]

@elly has been mucking around with a command in cargo for "build and install everything listed in all the package files". If we do this it ought to be on a pretty isolated VM, in a data center floating in orbit.

[elly]

You can currently achieve this with: 'cargo list | awk '{print $2}' | while read pkg; do cargo install --test $pkg; done' - I should write a command for cargo to do this on its own, though.

[brson]

Maybe we can integrate it with (or learn from) Travis CI

[graydon]

I agree we have to do more CI work; I think travis might be quite a bit of work. It only supports one host environment.

[brson]

We have most of the tools now to set up a pretty nice CI server completely in Rust and I want to desperately. By monitoring all the language-induced breakage going on in cargo we can start developing better tools to migrate versions.

When it fails to build a package, we could run it through a special compiler who's job is to upgrade the code. We will need a way to measure 'upgradedness' - it's possible we can do a lot of useful analysis and transformations without completely succeeding. Part of the process of developing the Rust language will be trying to improve the score of the upgrade compiler.

I'm putting this on 0.5 in the hopes that I can make some progress.

[brson]

Also, this is how I want to do CI for servo.

[catamorphism]

Bumping to 0.7

[catamorphism]

Not critical for 0.7. Nominating for milestone 4, well-covered.

[graydon]

accepted for well-covered milestone

[emberian]

triage bump

[brson]

This exists now: http://hiho.io/rust-ci/. We should encourage its development and eventually adopt it officially.

[catamorphism]

1.0, high

[emberian]

http://rust-ci.org is the new canonical URL.

[mrshu]

This seems to be quite an old issue and I am not really sure if it is still actually valid given the situation around rustpkg.

Maybe it might make sense to close this so that the 1.0 issue count can get lower?

[brson]

I'm OK closing since rust ci exists. Don't see this as 1.0 blocker.
On Feb 10, 2014 1:45 PM, "Marek Šuppa" [email protected] wrote:

This seems to be quite an old issue and I am not really sure if it is
still actually valid given the situation around rustpkg.

Maybe it might make sense to close this so that the 1.0 issue count can
get lower?

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/1453#issuecomment-34686295
.

[brson]

Closing. I think Rust CI's existence is enough to consider this done. Can follow up with more specific issues later.