docs: Use heading attributes to control the fragment.

.github/workflows/main.yml

@@ -232,7 +232,7 @@ jobs:
     - name: Install mdbook
       run: |
         mkdir mdbook
-        curl -Lf https://github.com/rust-lang/mdBook/releases/download/v0.4.27/mdbook-v0.4.27-x86_64-unknown-linux-gnu.tar.gz | tar -xz --directory=./mdbook
+        curl -Lf https://github.com/rust-lang/mdBook/releases/download/v0.4.31/mdbook-v0.4.31-x86_64-unknown-linux-gnu.tar.gz | tar -xz --directory=./mdbook
         echo `pwd`/mdbook >> $GITHUB_PATH
     - run: cd src/doc && mdbook build --dest-dir ../../target/doc
     - name: Run linkchecker.sh

src/doc/src/reference/build-scripts.md

@@ -34,7 +34,7 @@ The sections below describe how build scripts work, and the [examples
 chapter](build-script-examples.md) shows a variety of examples on how to write
 scripts.
 
-> Note: The [`package.build` manifest key](manifest.md#package-build) can be
+> Note: The [`package.build` manifest key](manifest.md#the-build-field) can be
 > used to change the name of the build script, or disable it entirely.
 
 ### Life Cycle of a Build Script
@@ -132,8 +132,7 @@ one detailed below.
   scripts.
 
 
-<a id="rustc-link-arg"></a>
+#### `cargo:rustc-link-arg=FLAG` {#rustc-link-arg}
-#### `cargo:rustc-link-arg=FLAG`
 
 The `rustc-link-arg` instruction tells Cargo to pass the [`-C link-arg=FLAG`
 option][link-arg] to the compiler, but only when building supported targets
@@ -143,8 +142,7 @@ linker script.
 
 [link-arg]: ../../rustc/codegen-options/index.md#link-arg
 
-<a id="rustc-link-arg-bin"></a>
+#### `cargo:rustc-link-arg-bin=BIN=FLAG` {#rustc-link-arg-bin}
-#### `cargo:rustc-link-arg-bin=BIN=FLAG`
 
 The `rustc-link-arg-bin` instruction tells Cargo to pass the [`-C
 link-arg=FLAG` option][link-arg] to the compiler, but only when building
@@ -152,8 +150,7 @@ the binary target with name `BIN`. Its usage is highly platform specific. It is
 to set a linker script or other linker options.
 
 
-<a id="rustc-link-arg-bins"></a>
+#### `cargo:rustc-link-arg-bins=FLAG` {#rustc-link-arg-bins}
-#### `cargo:rustc-link-arg-bins=FLAG`
 
 The `rustc-link-arg-bins` instruction tells Cargo to pass the [`-C
 link-arg=FLAG` option][link-arg] to the compiler, but only when building a
@@ -161,8 +158,7 @@ binary target. Its usage is highly platform specific. It is useful
 to set a linker script or other linker options.
 
 
-<a id="rustc-link-lib"></a>
+#### `cargo:rustc-link-lib=LIB` {#rustc-link-lib}
-#### `cargo:rustc-link-lib=LIB`
 
 The `rustc-link-lib` instruction tells Cargo to link the given library using
 the compiler's [`-l` flag][option-link]. This is typically used to link a
@@ -187,30 +183,26 @@ The optional `KIND` may be one of `dylib`, `static`, or `framework`. See the
 [FFI]: ../../nomicon/ffi.md
 
 
-<a id="rustc-link-arg-tests"></a>
+#### `cargo:rustc-link-arg-tests=FLAG` {#rustc-link-arg-tests}
-#### `cargo:rustc-link-arg-tests=FLAG`
 
 The `rustc-link-arg-tests` instruction tells Cargo to pass the [`-C
 link-arg=FLAG` option][link-arg] to the compiler, but only when building a
 tests target.
 
 
-<a id="rustc-link-arg-examples"></a>
+#### `cargo:rustc-link-arg-examples=FLAG` {#rustc-link-arg-examples}
-#### `cargo:rustc-link-arg-examples=FLAG`
 
 The `rustc-link-arg-examples` instruction tells Cargo to pass the [`-C
 link-arg=FLAG` option][link-arg] to the compiler, but only when building an examples
 target.
 
-<a id="rustc-link-arg-benches"></a>
+#### `cargo:rustc-link-arg-benches=FLAG` {#rustc-link-arg-benches}
-#### `cargo:rustc-link-arg-benches=FLAG`
 
 The `rustc-link-arg-benches` instruction tells Cargo to pass the [`-C
 link-arg=FLAG` option][link-arg] to the compiler, but only when building a benchmark
 target.
 
-<a id="rustc-link-search"></a>
+#### `cargo:rustc-link-search=[KIND=]PATH` {#rustc-link-search}
-#### `cargo:rustc-link-search=[KIND=]PATH`
 
 The `rustc-link-search` instruction tells Cargo to pass the [`-L`
 flag][option-search] to the compiler to add a directory to the library search
@@ -228,16 +220,14 @@ is fine).
 
 [option-search]: ../../rustc/command-line-arguments.md#option-l-search-path
 
-<a id="rustc-flags"></a>
+#### `cargo:rustc-flags=FLAGS` {#rustc-flags}
-#### `cargo:rustc-flags=FLAGS`
 
 The `rustc-flags` instruction tells Cargo to pass the given space-separated
 flags to the compiler. This only allows the `-l` and `-L` flags, and is
 equivalent to using [`rustc-link-lib`](#rustc-link-lib) and
 [`rustc-link-search`](#rustc-link-search).
 
-<a id="rustc-cfg"></a>
+#### `cargo:rustc-cfg=KEY[="VALUE"]` {#rustc-cfg}
-#### `cargo:rustc-cfg=KEY[="VALUE"]`
 
 The `rustc-cfg` instruction tells Cargo to pass the given value to the
 [`--cfg` flag][option-cfg] to the compiler. This may be used for compile-time
@@ -258,8 +248,7 @@ identifier, the value should be a string.
 [conditional compilation]: ../../reference/conditional-compilation.md
 [option-cfg]: ../../rustc/command-line-arguments.md#option-cfg
 
-<a id="rustc-env"></a>
+#### `cargo:rustc-env=VAR=VALUE` {#rustc-env}
-#### `cargo:rustc-env=VAR=VALUE`
 
 The `rustc-env` instruction tells Cargo to set the given environment variable
 when compiling the package. The value can be then retrieved by the [`env!`
@@ -279,8 +268,7 @@ Cargo][env-cargo].
 [env-macro]: ../../std/macro.env.html
 [env-cargo]: environment-variables.md#environment-variables-cargo-sets-for-crates
 
-<a id="rustc-cdylib-link-arg"></a>
+#### `cargo:rustc-cdylib-link-arg=FLAG` {#rustc-cdylib-link-arg}
-#### `cargo:rustc-cdylib-link-arg=FLAG`
 
 The `rustc-cdylib-link-arg` instruction tells Cargo to pass the [`-C
 link-arg=FLAG` option][link-arg] to the compiler, but only when building a
@@ -288,8 +276,7 @@ link-arg=FLAG` option][link-arg] to the compiler, but only when building a
 to set the shared library version or the runtime-path.
 
 
-<a id="cargo-warning"></a>
+#### `cargo:warning=MESSAGE` {#cargo-warning}
-#### `cargo:warning=MESSAGE`
 
 The `warning` instruction tells Cargo to display a warning after the build
 script has finished running. Warnings are only shown for `path` dependencies
@@ -334,8 +321,7 @@ FAQ](../faq.md#why-is-cargo-rebuilding-my-code).
 
 [`exclude` and `include` fields]: manifest.md#the-exclude-and-include-fields
 
-<a id="rerun-if-changed"></a>
+#### `cargo:rerun-if-changed=PATH` {#rerun-if-changed}
-#### `cargo:rerun-if-changed=PATH`
 
 The `rerun-if-changed` instruction tells Cargo to re-run the build script if
 the file at the given path has changed. Currently, Cargo only uses the
@@ -354,8 +340,7 @@ automatically handles whether or not the script itself needs to be recompiled,
 and of course the script will be re-run after it has been recompiled.
 Otherwise, specifying `build.rs` is redundant and unnecessary.
 
-<a id="rerun-if-env-changed"></a>
+#### `cargo:rerun-if-env-changed=NAME` {#rerun-if-env-changed}
-#### `cargo:rerun-if-env-changed=NAME`
 
 The `rerun-if-env-changed` instruction tells Cargo to re-run the build script
 if the value of an environment variable of the given name has changed.

src/doc/src/reference/manifest.md

@@ -53,7 +53,6 @@ Every manifest file consists of the following sections:
 * [`[profile]`](profiles.md) --- Compiler settings and optimizations.
 * [`[workspace]`](workspaces.md) --- The workspace definition.
 
-<a id="package-metadata"></a>
 ### The `[package]` section
 
 The first section in a `Cargo.toml` is `[package]`.
@@ -112,7 +111,6 @@ breaking change.
 [Resolver]: resolver.md
 [SemVer compatibility]: semver.md
 
-<a id="the-authors-field-optional"></a>
 #### The `authors` field
 
 The optional `authors` field lists in an array the people or organizations that are considered
@@ -135,7 +133,6 @@ user interface.
 > field cannot be changed or removed in already-published versions of a
 > package.
 
-<a id="the-edition-field-optional"></a>
 #### The `edition` field
 
 The `edition` key is an optional key that affects which [Rust Edition] your package
@@ -198,7 +195,6 @@ description = "A short description of my package"
 
 > **Note**: [crates.io] requires the `description` to be set.
 
-<a id="the-documentation-field-optional"></a>
 #### The `documentation` field
 
 The `documentation` field specifies a URL to a website hosting the crate's
@@ -326,7 +322,6 @@ categories = ["command-line-utilities", "development-tools::cargo-plugins"]
 > match one of the strings available at <https://crates.io/category_slugs>, and
 > must match exactly.
 
-<a id="the-workspace--field-optional"></a>
 #### The `workspace` field
 
 The `workspace` field can be used to configure the workspace that this package
@@ -347,8 +342,6 @@ table defined. That is, a crate cannot both be a root crate in a workspace
 
 For more information, see the [workspaces chapter](workspaces.md).
 
-<a id="package-build"></a>
-<a id="the-build-field-optional"></a>
 #### The `build` field
 
 The `build` field specifies a file in the package root which is a [build
@@ -368,7 +361,6 @@ The default is `"build.rs"`, which loads the script from a file named
 specify a path to a different file or `build = false` to disable automatic
 detection of the build script.
 
-<a id="the-links-field-optional"></a>
 #### The `links` field
 
 The `links` field specifies the name of a native library that is being linked
@@ -386,7 +378,6 @@ on Linux) may specify:
 links = "git2"
 ```
 
-<a id="the-exclude-and-include-fields-optional"></a>
 #### The `exclude` and `include` fields
 
 The `exclude` and `include` fields can be used to explicitly specify which
@@ -476,7 +467,6 @@ if any of those files change.
 
 [gitignore]: https://git-scm.com/docs/gitignore
 
-<a id="the-publish--field-optional"></a>
 #### The `publish` field
 
 The `publish` field can be used to prevent a package from being published to a
@@ -501,7 +491,6 @@ publish = ["some-registry-name"]
 If publish array contains a single registry, `cargo publish` command will use
 it when `--registry` flag is not specified.
 
-<a id="the-metadata-table-optional"></a>
 #### The `metadata` table
 
 Cargo by default will warn about unused keys in `Cargo.toml` to assist in
@@ -624,6 +613,17 @@ more detail.
         "#the-patch-section": "overriding-dependencies.html#the-patch-section",
         "#using-patch-with-multiple-versions": "overriding-dependencies.html#using-patch-with-multiple-versions",
         "#the-replace-section": "overriding-dependencies.html#the-replace-section",
+        "#package-metadata": "manifest.html#the-package-section",
+        "#the-authors-field-optional": "manifest.html#the-authors-field",
+        "#the-edition-field-optional": "manifest.html#the-edition-field",
+        "#the-documentation-field-optional": "manifest.html#the-documentation-field",
+        "#the-workspace--field-optional": "manifest.html#the-workspace-field",
+        "#package-build": "manifest.html#the-build-field",
+        "#the-build-field-optional": "manifest.html#the-build-field",
+        "#the-links-field-optional": "manifest.html#the-links-field",
+        "#the-exclude-and-include-fields-optional": "manifest.html#the-exclude-and-include-fields",
+        "#the-publish--field-optional": "manifest.html#the-publish-field",
+        "#the-metadata-table-optional": "manifest.html#the-metadata-table",
     };
     var target = fragments[window.location.hash];
     if (target) {

src/doc/src/reference/semver.md

@@ -112,8 +112,7 @@ after it has been modified, and an example usage of the code that could appear
 in another project. In a minor change, the example usage should successfully
 build with both the before and after versions.
 
-<a id="item-remove"></a>
+### Major: renaming/moving/removing any public items {#item-remove}
-### Major: renaming/moving/removing any public items
 
 The absence of a publicly exposed [item][items] will cause any uses of that item to
 fail to compile.
@@ -145,8 +144,7 @@ Mitigating strategies:
 * Mark renamed items as [deprecated], and use a [`pub use`] item to re-export
   to the old name.
 
-<a id="item-new"></a>
+### Minor: adding new public items {#item-new}
-### Minor: adding new public items
 
 Adding new, public [items] is a minor change.
 
@@ -206,8 +204,7 @@ This is not considered a major change because conventionally glob imports are
 a known forwards-compatibility hazard. Glob imports of items from external
 crates should be avoided.
 
-<a id="struct-add-private-field-when-public"></a>
+### Major: adding a private struct field when all current fields are public {#struct-add-private-field-when-public}
-### Major: adding a private struct field when all current fields are public
 
 When a private field is added to a struct that previously had all public fields,
 this will break any code that attempts to construct it with a [struct literal].
@@ -241,8 +238,7 @@ Mitigation strategies:
   a struct to prevent users from using struct literal syntax, and instead
   provide a constructor method and/or [Default] implementation.
 
-<a id="struct-add-public-field-when-no-private"></a>
+### Major: adding a public field when no private field exists {#struct-add-public-field-when-no-private}
-### Major: adding a public field when no private field exists
 
 When a public field is added to a struct that has all public fields, this will
 break any code that attempts to construct it with a [struct literal].
@@ -276,8 +272,7 @@ Mitigation strategies:
   a struct to prevent users from using struct literal syntax, and instead
   provide a constructor method and/or [Default] implementation.
 
-<a id="struct-private-fields-with-private"></a>
+### Minor: adding or removing private fields when at least one already exists {#struct-private-fields-with-private}
-### Minor: adding or removing private fields when at least one already exists
 
 It is safe to add or remove private fields from a struct when the struct
 already has at least one private field.
@@ -335,8 +330,7 @@ fn main() {
 }
 ```
 
-<a id="struct-tuple-normal-with-private"></a>
+### Minor: going from a tuple struct with all private fields (with at least one field) to a normal struct, or vice versa {#struct-tuple-normal-with-private}
-### Minor: going from a tuple struct with all private fields (with at least one field) to a normal struct, or vice versa
 
 Changing a tuple struct to a normal struct (or vice-versa) is safe if all
 fields are private.
@@ -367,8 +361,7 @@ fn main() {
 This is safe because existing code cannot use a [struct literal] to construct
 it, nor match its contents.
 
-<a id="enum-variant-new"></a>
+### Major: adding new enum variants (without `non_exhaustive`) {#enum-variant-new}
-### Major: adding new enum variants (without `non_exhaustive`)
 
 It is a breaking change to add a new enum variant if the enum does not use the
 [`#[non_exhaustive]`][non_exhaustive] attribute.
@@ -404,8 +397,7 @@ Mitigation strategies:
 * When introducing the enum, mark it as [`#[non_exhaustive]`][non_exhaustive]
   to force users to use [wildcard patterns] to catch new variants.
 
-<a id="enum-fields-new"></a>
+### Major: adding new fields to an enum variant {#enum-fields-new}
-### Major: adding new fields to an enum variant
 
 It is a breaking change to add new fields to an enum variant because all
 fields are public, and constructors and matching will fail to compile.
@@ -457,8 +449,7 @@ Mitigation strategies:
   }
   ```
 
-<a id="trait-new-item-no-default"></a>
+### Major: adding a non-defaulted trait item {#trait-new-item-no-default}
-### Major: adding a non-defaulted trait item
 
 It is a breaking change to add a non-defaulted item to a trait. This will
 break any implementors of the trait.
@@ -490,8 +481,7 @@ Mitigation strategies:
 * When introducing the trait, use the [sealed trait] technique to prevent
   users outside of the crate from implementing the trait.
 
-<a id="trait-item-signature"></a>
+### Major: any change to trait item signatures {#trait-item-signature}
-### Major: any change to trait item signatures
 
 It is a breaking change to make any change to a trait item signature. This can
 break external implementors of the trait.
@@ -530,8 +520,7 @@ Mitigation strategies:
 * When introducing the trait, use the [sealed trait] technique to prevent
   users outside of the crate from implementing the trait.
 
-<a id="trait-new-default-item"></a>
+### Possibly-breaking: adding a defaulted trait item {#trait-new-default-item}
-### Possibly-breaking: adding a defaulted trait item
 
 It is usually safe to add a defaulted trait item. However, this can sometimes
 cause a compile error. For example, this can introduce an ambiguity if a
@@ -581,8 +570,7 @@ Mitigation strategies:
   to require downstream users to add [disambiguation syntax] to select the
   correct function when updating the dependency.
 
-<a id="trait-object-safety"></a>
+### Major: adding a trait item that makes the trait non-object safe {#trait-object-safety}
-### Major: adding a trait item that makes the trait non-object safe
 
 It is a breaking change to add a trait item that changes the trait to not be
 [object safe].
@@ -616,8 +604,7 @@ fn main() {
 It is safe to do the converse (making a non-object safe trait into a safe
 one).
 
-<a id="trait-new-parameter-no-default"></a>
+### Major: adding a type parameter without a default {#trait-new-parameter-no-default}
-### Major: adding a type parameter without a default
 
 It is a breaking change to add a type parameter without a default to a trait.
 
@@ -643,8 +630,7 @@ impl Trait for Foo {}  // Error: missing generics
 Mitigating strategies:
 * See [adding a defaulted trait type parameter](#trait-new-parameter-default).
 
-<a id="trait-new-parameter-default"></a>
+### Minor: adding a defaulted trait type parameter {#trait-new-parameter-default}
-### Minor: adding a defaulted trait type parameter
 
 It is safe to add a type parameter to a trait as long as it has a default.
 External implementors will use the default without needing to specify the
@@ -669,8 +655,7 @@ struct Foo;
 impl Trait for Foo {}
 ```
 
-<a id="impl-item-new"></a>
+### Possibly-breaking change: adding any inherent items {#impl-item-new}
-### Possibly-breaking change: adding any inherent items
 
 Usually adding inherent items to an implementation should be safe because
 inherent items take priority over trait items. However, in some cases the
@@ -719,8 +704,7 @@ Mitigation strategies:
   to require downstream users to add [disambiguation syntax] to select the
   correct function when updating the dependency.
 
-<a id="generic-bounds-tighten"></a>
+### Major: tightening generic bounds {#generic-bounds-tighten}
-### Major: tightening generic bounds
 
 It is a breaking change to tighten generic bounds on a type since this can
 break users expecting the looser bounds.
@@ -749,8 +733,7 @@ fn main() {
 }
 ```
 
-<a id="generic-bounds-loosen"></a>
+### Minor: loosening generic bounds {#generic-bounds-loosen}
-### Minor: loosening generic bounds
 
 It is safe to loosen the generic bounds on a type, as it only expands what is
 allowed.
@@ -779,8 +762,7 @@ fn main() {
 }
 ```
 
-<a id="generic-new-default"></a>
+### Minor: adding defaulted type parameters {#generic-new-default}
-### Minor: adding defaulted type parameters
 
 It is safe to add a type parameter to a type as long as it has a default. All
 existing references will use the default without needing to specify the
@@ -810,8 +792,7 @@ fn main() {
 }
 ```
 
-<a id="generic-generalize-identical"></a>
+### Minor: generalizing a type to use generics (with identical types) {#generic-generalize-identical}
-### Minor: generalizing a type to use generics (with identical types)
 
 A struct or enum field can change from a concrete type to a generic type
 parameter, provided that the change results in an identical type for all
@@ -840,8 +821,7 @@ fn main() {
 because existing uses of `Foo` are shorthand for `Foo<u8>` which yields the
 identical field type.
 
-<a id="generic-generalize-different"></a>
+### Major: generalizing a type to use generics (with possibly different types) {#generic-generalize-different}
-### Major: generalizing a type to use generics (with possibly different types)
 
 Changing a struct or enum field from a concrete type to a generic type
 parameter can break if the type can change.
@@ -866,8 +846,7 @@ fn main() {
 }
 ```
 
-<a id="generic-more-generic"></a>
+### Minor: changing a generic type to a more generic type {#generic-more-generic}
-### Minor: changing a generic type to a more generic type
 
 It is safe to change a generic type to a more generic one. For example, the
 following adds a generic parameter that defaults to the original type, which
@@ -894,8 +873,7 @@ fn main() {
 }
 ```
 
-<a id="fn-change-arity"></a>
+### Major: adding/removing function parameters {#fn-change-arity}
-### Major: adding/removing function parameters
 
 Changing the arity of a function is a breaking change.
 
@@ -924,8 +902,7 @@ Mitigating strategies:
   with the builder pattern. This allows new fields to be added to the struct
   in the future.
 
-<a id="fn-generic-new"></a>
+### Possibly-breaking: introducing a new function type parameter {#fn-generic-new}
-### Possibly-breaking: introducing a new function type parameter
 
 Usually, adding a non-defaulted type parameter is safe, but in some
 cases it can be a breaking change:
@@ -955,8 +932,7 @@ other ways) that this breakage is usually acceptable. One should take into
 account how likely it is that the function in question is being called with
 explicit type arguments.
 
-<a id="fn-generalize-compatible"></a>
+### Minor: generalizing a function to use generics (supporting original type) {#fn-generalize-compatible}
-### Minor: generalizing a function to use generics (supporting original type)
 
 The type of a parameter to a function, or its return value, can be
 *generalized* to use generics, including by introducing a new type parameter,
@@ -1053,8 +1029,7 @@ fn main() {
 }
 ```
 
-<a id="fn-generalize-mismatch"></a>
+### Major: generalizing a function to use generics with type mismatch {#fn-generalize-mismatch}
-### Major: generalizing a function to use generics with type mismatch
 
 It is a breaking change to change a function parameter or return type if the
 generic type constrains or changes the types previously allowed. For example,
@@ -1081,8 +1056,7 @@ fn main() {
 }
 ```
 
-<a id="fn-unsafe-safe"></a>
+### Minor: making an `unsafe` function safe {#fn-unsafe-safe}
-### Minor: making an `unsafe` function safe
 
 A previously `unsafe` function can be made safe without breaking code.
 
@@ -1122,8 +1096,7 @@ Making a previously `unsafe` associated function or method on structs / enums
 safe is also a minor change, while the same is not true for associated
 function on traits (see [any change to trait item signatures](#trait-item-signature)).
 
-<a id="attr-no-std-to-std"></a>
+### Major: switching from `no_std` support to requiring `std` {#attr-no-std-to-std}
-### Major: switching from `no_std` support to requiring `std`
 
 If your library specifically supports a [`no_std`] environment, it is a
 breaking change to make a new release that requires `std`.
@@ -1158,8 +1131,7 @@ Mitigation strategies:
   optionally enables `std` support, and when the feature is off, the library
   can be used in a `no_std` environment.
 
-<a id="attr-adding-non-exhaustive"></a>
+### Major: adding `non_exhaustive` to an existing enum, variant, or struct with no private fields {#attr-adding-non-exhaustive}
-### Major: adding `non_exhaustive` to an existing enum, variant, or struct with no private fields
 
 Making items [`#[non_exhaustive]`][non_exhaustive] changes how they may
 be used outside the crate where they are defined:
@@ -1243,8 +1215,7 @@ Mitigation strategies:
 
 ## Tooling and environment compatibility
 
-<a id="env-new-rust"></a>
+### Possibly-breaking: changing the minimum version of Rust required {#env-new-rust}
-### Possibly-breaking: changing the minimum version of Rust required
 
 Introducing the use of new features in a new release of Rust can break
 projects that are using older versions of Rust. This also includes using new
@@ -1271,8 +1242,7 @@ Mitigation strategies:
   mechanism for new features. These are currently unstable and only available
   in the nightly channel.
 
-<a id="env-change-requirements"></a>
+### Possibly-breaking: changing the platform and environment requirements {#env-change-requirements}
-### Possibly-breaking: changing the platform and environment requirements
 
 There is a very wide range of assumptions a library makes about the
 environment that it runs in, such as the host platform, operating system
@@ -1292,8 +1262,7 @@ Mitigation strategies:
 * Document the platforms and environments you specifically support.
 * Test your code on a wide range of environments in CI.
 
-<a id="new-lints"></a>
+### Minor: introducing new lints {#new-lints}
-### Minor: introducing new lints
 
 Some changes to a library may cause new lints to be triggered in users of that library.
 This should generally be considered a compatible change.
@@ -1348,8 +1317,7 @@ Mitigating strategies:
 
 ### Cargo
 
-<a id="cargo-feature-add"></a>
+#### Minor: adding a new Cargo feature {#cargo-feature-add}
-#### Minor: adding a new Cargo feature
 
 It is usually safe to add new [Cargo features]. If the feature introduces new
 changes that cause a breaking change, this can cause difficulties for projects
@@ -1371,8 +1339,7 @@ consequences of enabling the feature.
 std = []
 ```
 
-<a id="cargo-feature-remove"></a>
+#### Major: removing a Cargo feature {#cargo-feature-remove}
-#### Major: removing a Cargo feature
 
 It is usually a breaking change to remove [Cargo features]. This will cause
 an error for any project that enabled the feature.
@@ -1398,8 +1365,7 @@ Mitigation strategies:
   functionality. Document that the feature is deprecated, and remove it in a
   future major SemVer release.
 
-<a id="cargo-feature-remove-another"></a>
+#### Major: removing a feature from a feature list if that changes functionality or public items {#cargo-feature-remove-another}
-#### Major: removing a feature from a feature list if that changes functionality or public items
 
 If removing a feature from another feature, this can break existing users if
 they are expecting that functionality to be available through that feature.
@@ -1420,8 +1386,7 @@ default = []  # This may cause packages to fail if they are expecting std to be
 std = []
 ```
 
-<a id="cargo-remove-opt-dep"></a>
+#### Possibly-breaking: removing an optional dependency {#cargo-remove-opt-dep}
-#### Possibly-breaking: removing an optional dependency
 
 Removing an optional dependency can break a project using your library because
 another project may be enabling that dependency via [Cargo features].
@@ -1454,8 +1419,7 @@ Mitigation strategies:
   optional dependencies necessary to implement "networking". Then document the
   "networking" feature.
 
-<a id="cargo-change-dep-feature"></a>
+#### Minor: changing dependency features {#cargo-change-dep-feature}
-#### Minor: changing dependency features
 
 It is usually safe to change the features on a dependency, as long as the
 feature does not introduce a breaking change.
@@ -1475,8 +1439,7 @@ rand = { version = "0.7.3", features = ["small_rng"] }
 rand = "0.7.3"
 ```
 
-<a id="cargo-dep-add"></a>
+#### Minor: adding dependencies {#cargo-dep-add}
-#### Minor: adding dependencies
 
 It is usually safe to add new dependencies, as long as the new dependency
 does not introduce new requirements that result in a breaking change.

src/doc/src/reference/workspaces.md

@@ -56,7 +56,6 @@ version = "0.1.0"    # the current version, obeying semver
 authors = ["Alice <a@example.com>", "Bob <b@example.com>"]
 ```
 
-<a id="virtual-manifest"></a>
 #### Virtual workspace
 
 Alternatively, a `Cargo.toml` file can be created with a `[workspace]` section
@@ -253,3 +252,17 @@ if that makes sense for the tool in question.
 [specifying-dependencies]: specifying-dependencies.md
 [features]: features.md
 [inheriting-a-dependency-from-a-workspace]: specifying-dependencies.md#inheriting-a-dependency-from-a-workspace
+
+<script>
+(function() {
+    var fragments = {
+        "#virtual-manifest": "workspaces.html#virtual-workspace",
+    };
+    var target = fragments[window.location.hash];
+    if (target) {
+        var url = window.location.toString();
+        var base = url.substring(0, url.lastIndexOf('/'));
+        window.location.replace(base + "/" + target);
+    }
+})();
+</script>