Before:
```
Testing ["rustc_interface"] stage0 (aarch64-unknown-linux-gnu -> aarch64-unknown-linux-gnu)
```
After:
```
Testing {rustc_interface} stage0 (aarch64-unknown-linux-gnu -> aarch64-unknown-linux-gnu)
```
Note there is a slight consistency between `build` and `test`: The
former doesn't print "compiler artifacts". It would be annoying to fix
and doesn't hurt anything, so I left it be.
```
; x t rustc_interface --stage 0 --dry-run
Testing {rustc_interface} stage0 (aarch64-unknown-linux-gnu -> aarch64-unknown-linux-gnu)
; x b rustc_interface --stage 0 --dry-run
Building {rustc_interface} stage0 compiler artifacts (aarch64-unknown-linux-gnu -> aarch64-unknown-linux-gnu)
```
copy doc output files by format
This pr provides copying doc outputs by checking output format without removing output directory on each trigger.
Resolves#103785
This allows open() to be called from other places in bootstrap (I need
this for Ferrocene), and it simplifies the callers by moving the
"was_invoked_explicitly" check into the function.
Support excluding the generation of the standalone docs
For Ferrocene we need to exclude the generation of the standalone docs (which include the index page, which we want to replace with our own), but with the way bootstrap is currently implemented that proved not possible. This PR aims to support that.
The first problem is that the `doc::Standalone` step did two things: it generated the "standalone" documentation (which includes the index page and all the pages at the root of the documentation tree), but it also generated some files like `rust.css` and `version_info.html` that other step like `doc::TheBook` required. This meant generating the book required generating the index page, which made disabling the index page generation problematic.
The approach I took to fix the first problem is to split the step into `doc::Standalone` and `doc::SharedAssets`, with `doc::TheBook` now depending on `doc::SharedAssets`.
The second problem is that disabling the `doc::Standalone` proved to be tricky due to its path, `src/doc`. The path is accurate, as the source files for that step are `src/doc/*.md`. The problem is, bootstrap treats `--exclude` as a *suffix*, and so it also excluded the Cargo book whose source lives at `src/tools/cargo/src/doc`.
The approach I took to fix the second problem is to add the `standalone` path in addition to `src/doc`, so that you can pass `--exclude standalone`. I'm not fully happy with the solution, and the other idea I had was to just move the standalone docs source code to `src/doc/standalone`. I feel that second approach is cleaner, but also requires more changes and might require more consensus.
This PR is best reviewed commit-by-commit.
r? `@jyn514`
Before this commit, the path for the doc::Standalone step was "src/doc",
which is accurate as the standalone docs source files live at the root
of the "src/doc" directory tree.
Unfortunately, that caused bad interactions when trying to exclude it
with `--exclude src/doc`. When an exclusion is passed to bootstrap, it
will exclude all steps whose path *ends with* the exclusion, which
results in the Cargo book (src/tools/cargo/src/doc) to also be excluded.
To work around this problem, this commit adds the "standalone" path as
an alternate path for doc::Standalone, allowing `--exclude standalone`
to work without side effects.
Before this commit, the step to generate the standalone docs (which
included the index page and other HTML files at the root of the
documentation) was bundled with the code copying files needed by
multiple pieces of documentation. This means it wasn't possible to avoid
generating the standalone docs.
This commit splits the step into two, allowing the standalone docs
generation to be excluded while still building the rest of the docs.
Distribute json doc
# Overview
We add a new component, `rust-json-docs`, to distribute the JSON version of rustdoc's output for public compiler crates (i.e. `std`, `alloc`, `proc_macro`, `core` and `test`).
As discussed in #101383, we do not bundle this up as part of the existing `rust-docs` component since `rustdoc`'s JSON format is still unstable.
# Open questions / Doubts
I tried my best, but I never touched this codebase and I couldn't find much documentation on how `dist` works - I pattern-matched existing code, which might have led to some non-sensical choices in the eyes of people more familiar with the codebase. In particular, I am not sure if my choice of adding a new config flag is appropriate or if the decision to build/not build the JSON docs is more appropriately gated by one of the existing flags.
Any suggestion is more than welcome.
Closes#101383
This was originally added so those doc pages could use the same font
files, but it turns out to be fragile. And those doc pages are just
stubs that link to other pages, so they don't need fancy fonts.
bootstrap: Allow building individual crates
This aims to be as unintrusive as possible, but did still require adding a new `tail_args` field to all `Rustc` and `Std` steps.
New library and compiler crates are added to the sysroot as they are built, since it's useful to have e.g. just alloc and not std.
Fixes https://github.com/rust-lang/rust/issues/44293.
Add `x {check,build,doc} {compiler,library}` aliases.
While working on https://github.com/rust-lang/rust/pull/95503, I realized that it will interfere with existing command lines:
Currently people run `x build library/std` expecting it to "add all library crates to the sysroot",
but after that change, it will *only* build `libstd` and its dependencies (and add them to the sysroot), not libtest or libproc_macro.
That will work for local testing in most cases, but could be confusing. Even if not, though, I think `x build library` is more clear about what actually happens than the current `x build library/std`.
The intended end goal is something like:
- For check/build/doc, we have library + compiler aliases, which correspond to basically "most possible" for that piece. This is the intended path of entry (rather than library/test or similar as today) for when you just want the thing to work -- for example, getting a compiler that is "crates.io-compatible" would be roughly `x.py build library`). #95504
- Specific crate invocations build up to that crate, which means that if you don't care about tests you probably want x.py build library/proc_macro or library/std for faster build times. #95503
Note that this is already implemented today for the `doc` command and seems to work pretty well in practice.
I plan to change the dev-guide and various instructions in the README to `build library` once this is merged.
`@rustbot` label +A-rustbuild
Rustdoc doesn't require the build artifacts to generate the docs, and
especially in the case of rustc, it greatly increases the time needed to
run the build.
- Statically ensure that only the top_stage of a tool is documented
If another part of rustbuild tried to document a different stage, it
would run into errors because `check::Rustc` unconditionally uses the
top stage.
- Try building rustc instead of checking to avoid duplicate artifacts
Tries to workaround the following error:
```
error[E0464]: multiple matching crates for `rustc_ast`
--> src/librustdoc/lib.rs:40:1
|
40 | extern crate rustc_ast;
| ^^^^^^^^^^^^^^^^^^^^^^^
|
= note: candidates:
crate `rustc_ast`: /checkout/obj/build/x86_64-unknown-linux-gnu/stage2/lib/rustlib/x86_64-unknown-linux-gnu/lib/librustc_ast-6d7c193782263d89.rlib
crate `rustc_ast`: /checkout/obj/build/x86_64-unknown-linux-gnu/stage2/lib/rustlib/x86_64-unknown-linux-gnu/lib/librustc_ast-e5d09eda5beb759c.rmeta
```
While working on https://github.com/rust-lang/rust/pull/95503,
I realized that this will interfere with existing command lines:
Currently people run `x build library/std` expecting it to be added to the sysroot,
but after that change, it will *only* build `libstd` without making it available
for the toolchain.
It's debatable whether that's a breaking change that will be accepted; if so, this PR is absolutely
necessary to make sure there's a command for "build the standard library and add it to the sysroot".
Even if not, though, I think `x build library` is more clear about what actually happens than the
current `x build library/std`.
For consistency, also add support for `compiler` and all other command variants. Note that `doc
compiler` was already supported, so in a sense this is just fixing an existing inconsistency.
I plan to change the dev-guide and various instructions in the README to `build library` once this is merged.
Joshua Nelson
clubby789
Alex Huang
Pietro Albini
Michael Goulet
Lukas Markeffsky
Joshua Nelson
Nixon Enraght-Moony
Dylan DPC
ozkanonur
Matthias Krüger
viandoxdev
bors
Pietro Albini
Oli Scherer
Josh Triplett
Luca Palmieri
Jacob Hoffman-Andrews
Guillaume Gomez
flip1995
Guillaume Gomez
Ralf Jung
InfRandomness
Joshua Nelson
Eric Huss