Add initial version of snapshot tests to bootstrap
When making any changes to bootstrap (steps), it is very difficult to realize how does it affect various common bootstrap commands, and if everything still works as we expect it to. We are far away from having actual end-to-end tests, but what we could at least do is have a way of testing what steps does bootstrap execute in dry run mode. Now, we already have something like this in `src/bootstrap/src/core/builder/tests.rs`, however that is quite limited, because it only checks executed steps for a specific impl of `Step` and it does not consider step order.
Recently, when working on what I thought was one of the simplest possible step untanglings in bootstrap (https://github.com/rust-lang/rust/pull/142357), I ran into errors in tests that were quite hard to debug. Partly also because the current staging test diffs are multiline and use `Debug` output, so it's quite difficult for me to make sense of them.
In this PR, I introduce `insta`, which allows writing snapshot tests in a very simple way. With it, I want to allow writing tests that will clearly show us what is going on during bootstrap execution, and then write golden tests for `build/check/test` stage `0/1/2` for compiler/std/tools etc., to make sure that we don't regress something, and also to help with [#t-infra/bootstrap > Proposal to cleanup stages and steps after the redesign](https://rust-lang.zulipchat.com/#narrow/channel/326414-t-infra.2Fbootstrap/topic/Proposal.20to.20cleanup.20stages.20and.20steps.20after.20the.20redesign/with/523488806), to help avoid a situation where we would (again) have to make a flurry of staging changes because of unexpected consequences.
In the snapshot tests, we currently render the build of rustc, std and LLVM. Currently I render the executed steps using downcasting, which is not super pretty, but it allows us to make the test rendering localized in one place, and it's IMO enough for now.
I implemented only a single test using the new machinery. Maybe if you take a look at it, you will understand why 😆 Bootstrap currently does some peculiar things, such as running a stage 0 std step (even though stage 0 std no longer exists) and running the Rustc stage 0 -> 1 step twice, once with a single crates, once with all rustc crates. So I think that even with this single step, there will be a bunch of things to fix in the near future...
The way we currently prepare the Config test fixtures is far from ideal, this is something I think ``@Shourya742`` could work on as a part of their GSoC project (remove as much command execution from Config construction as possible, actually run bootstrap on a temporary directory instead of running it on the rustc checkout, create a Builder-like API for creating the Config test fixtures).
r? ``@jieyouxu``
Assorted bootstrap cleanups (step 2)
Very small improvements designed towards making bootstrap tests less hacky/special, and towards making it possible to run bootstrap tests in parallel.
Best reviewed commit by commit.
r? ``@jieyouxu``
Add bootstrap option to compile a tool with features
Add an option to specify which features to build a tool with, e.g. it will be useful to build Miri with tracing enabled:
```toml
tool-config.miri.features = ["tracing"]
```
See [this Zulip thread](https://rust-lang.zulipchat.com/#narrow/channel/326414-t-infra.2Fbootstrap/topic/Passing.20--features.20to.20Miri.20build.20using.20.2E.2Fx.2Epy/with/523564773) for the options considered. If the final decision will be different than what I wrote now, I will update the code as needed. The reason why the option is `tool-config.miri.features` instead of something like `tool-features.miri` is to possibly allow adding more tool-specific configurations in the future.
I didn't do any validation of the keys of the `tool-config` hashmap, since I saw that no validation is done on the `tools` hashset either.
I don't like much the fact that features can be chosen by various places of the codebase: `Step`s can have some fixed `extra_features`, `prepare_tool_cargo` will add features depending on some bootstrapping options, and the newly added option can also contribute features to tools. However I think it is out of scope of this PR to try to refactor all of that (if it even is refactorable), so I left a comment in the codebase explaining all of the sources of features I could find.
Assorted bootstrap cleanups (step 1)
Now that the stage0 redesign has landed, we can finally start cleaning up many things in bootstrap, and lord knows it deserves it! I plan to send many PRs once I figure out an incremental way forward, this is the first one of them. It doesn't actually change anything, just renames stuff and adds more documentation, but the rename is bitrotty, so I wanted to push the PR eagerly.
r? `@jieyouxu`
Add tracing import to execution context
In https://github.com/rust-lang/rust/pull/141909, we missed adding the trace_cmd import in the execution context module. This PR fixes that. Additionally, we are updating the mingw-check-2 check command to include BOOTSTRAP_TRACING=1 to help ensure we don't miss such cases in future PRs.
r? `@Kobzol`
Implement `//@ needs-target-std` compiletest directive
Closesrust-lang/rust#141863.
Needed to unblock rust-lang/rust#139244 and rust-lang/rust#141856.
### Summary
This PR implements a `//@ needs-target-std` compiletest directive that gates test execution based on whether the target supports std or not. For some cases, this should be preferred over e.g. some combination of `//@ ignore-none`, `//@ ignore-nvptx` and more[^none-limit].
### Implementation limitation
Unfortunately, since there is currently [no reliable way to determine from metadata whether a given target supports std or not](https://github.com/rust-lang/rust/issues/142296), we have to resort to a hack. Bootstrap currently determines whether or not a target supports std by a naive target tuple substring comparison: a target supports std if its target tuple does *not* contain one of `["-none", "nvptx", "switch"]` substrings. This PR simply pulls that hack out into `build_helpers` to avoid reimplementing the same hack in compiletest, and uses that logic to inform `//@ needs-target-std`.
### Auxiliary changes
This PR additionally changes a few run-make tests to use `//@ needs-target-std` over an inconsistent combination of target-based `ignore`s. This should help with rust-lang/rust#139244.
---
r? bootstrap
[^none-limit]: Notably, `target_os = "none"` is **not** a sufficient condition for "target does not support std"
Host is the machine where bootstrap runs, and this field represents the target of the (host) stage0/beta compiler. This is much clearer than `build`, which also conflicts with the `Build` struct, which is stored under the name `build` inside `Builder` (lol).
Add central execution context to bootstrap
This PR continues the effort toward command centralization as outlined in https://github.com/rust-lang/rust/issues/126819. It introduces a centralized execution context through which all commands will be executed. Previously, centralization was limited to build methods; this PR extends it to the `config` module and updates the remaining methods accordingly.
Best reviewed commit by commit.
r? ``@Kobzol``
To centralize this hack in one place with a backlink to the issue
tracking this hack, as this logic is also needed by compiletest to
implement a `//@ needs-target-std` directive.
