This tool now scans for cargo dependencies and includes any important looking license files.

We do this because cargo package metadata is not sufficient - the Apache-2.0 license says you have to include any NOTICE file, for example. And authors != copyright holders (cargo has the former, we must include the latter).

This format works better with large amounts of structured data.

We also mark which deps are in the stdlib

I can't find a way to derive rinja::Template for Node - I think because it is a recursive type. So I rendered it manually using html_escape.

…ributes

…ling tests

Co-authored-by: Amanieu d'Antras <[email protected]>

The object crate was recently updated to recognize the 32-bit SPARC
ELF targets EM_SPARC and EM_SPARC32PLUS, so the proper architecture
for 32-bit SPARC can now be set in rustc_codegen_ssa.

…res, r=Amanieu

Add implied target features to target_feature attribute

See [zulip](https://rust-lang.zulipchat.com/#narrow/stream/208962-t-libs.2Fstdarch/topic/Why.20would.20target-feature.20include.20implied.20features.3F) for some context.  Adds implied target features, e.g. `#[target_feature(enable = "avx2")]` acts like `#[target_feature(enable = "avx2,avx,sse4.2,sse4.1...")]`.  Fixes rust-lang#128125, fixes rust-lang#128426

The implied feature sets are taken from [the rust reference](https://doc.rust-lang.org/reference/attributes/codegen.html?highlight=target-fea#x86-or-x86_64), there are certainly more features and targets to add.

Please feel free to reassign this to whoever should review it.

r? ``@Amanieu``

impl `Default` for collection iterators that don't already have it

There is a pretty strong precedent for implementing `Default` for collection iterators, and this does so for some where this implementation was missed.

I don't think this needs a separate ACP (since this precedent already exists, and these feel like they were just missed), however, it *will* need an FCP since these implementations are instantly stable.

…ndencies-to-copyright-file, r=Kobzol

Change generate-copyright to generate HTML, with cargo dependencies included

`x.py run generate-copyright` now produces `build/COPYRIGHT.html`. This includes a new format for in-tree dependencies, and also adds out-of-tree cargo dependencies.

After consulting expert opinion, I have elected to include every top-level:

* `*NOTICE*`
* `*AUTHOR*`
* `*LICENSE*`
* `*LICENCE*`, and
* `*COPYRIGHT*` file I can find - case-insensitive.

This is because the cargo package metadata's `author` field is not a list of copyright holders and does not meet the requirements of the Apache-2.0 license (which says you must include a NOTICE file with the binary if one was supplied by the author) nor the MIT license (which says you must include 'the above copyright notice').

I believe it would be appropriate to include this file with every Rust release, in order to do an even better job of appropriately recognising the efforts of the authors of the first-party and third-party libraries we are using here.

The output includes something like 524 copies of the Apache-2.0 text because they are not all identical. I think I count about 50 different variations by shasum - some differ in whitespace, while some have the boilerplate block at the bottom erroneously modified (don't modify the copy in the license, modify the copy you paste into your own source code!). Running `gzip` on the HTML file largely makes this problem go away, and the average browser is far happier with a ~6 MiB HTML file than the average Markdown viewer is with a ~6 MiB markdown file. But, if someone wants to, do they could submit a follow-up which de-dups the license text files and adds back-links to earlier identical copies (for some value of 'identical copy').

```console
$ xpy run generate-copyright
$ cd build
$ gzip -c COPYRIGHT.html > COPYRIGHT.gz
$ xz -c COPYRIGHT.html > COPYRIGHT.xz
$ ls -lh COPYRIGHT.*
-rw-r--r--  1 jonathan  staff   241K 29 Jul 17:19 COPYRIGHT.gz
-rw-r--r--@ 1 jonathan  staff   6.6M 29 Jul 11:30 COPYRIGHT.html
-rw-r--r--  1 jonathan  staff    59K 29 Jul 17:19 COPYRIGHT.xz
```

Here's an example [COPYRIGHT.gz](https://github.com/user-attachments/files/16416147/COPYRIGHT.gz).

codegen: better centralize function declaration attribute computation

For some reason, the codegen backend has two functions that compute which attributes a function declaration gets: `apply_attrs_llfn` and `attributes::from_fn_attrs`. They are called in different places, on entirely different layers of abstraction.

To me the code seems cleaner if we centralize this entirely in `apply_attrs_llfn`, so that's what this PR does.

…trochenkov

make `import.vis` is immutable

r? `@petrochenkov`

Integrate crlf directly into related test file instead via of .gitattributes

resolves rust-lang#128708

This PR seeks to resolve a contributor papercut when using jj to manage the git repo locally which does not support .gitattributes. It does so by integrating the crlf characters directly into the related test and disabling Git's end of line normalization logic across platforms for that specific file, instead of configuring git to always check out the files with alternative eol characters.

related documentation: https://git-scm.com/docs/gitattributes#Documentation/gitattributes.txt-Unset-1

rustc_codegen_ssa: Set architecture for object crate for 32-bit SPARC

The `object` crate was recently updated to recognize the 32-bit SPARC ELF targets `EM_SPARC` and `EM_SPARC32PLUS`, so the proper architecture for 32-bit SPARC can now be set in `rustc_codegen_ssa`.

r? nagisa

…piler-errors

unused_parens: do not lint against parens around &raw

Requested by `@tmandry` in rust-lang#127679: with `&raw` one somewhat regularly has to write code like `(&raw const (*myptr).field).method()`, so parentheses around the expression are often required. To avoid churn between adding and removing parentheses as method calls appear and disappear, the proposal was made to silence the lint for unnecessary parentheses around `&raw` expressions. This PR implements that.