doc/README: More anchors to link to

Converts that long list into headers such that they can be linked to.
This commit is contained in:
Silvan Mosberger 2024-10-10 23:36:35 +02:00
parent e49f3574ab
commit 3a26957d8e

View File

@ -176,126 +176,138 @@ When needed, each convention explain why it exists, so you can make a decision w
Note that these conventions are about the **structure** of the manual (and its source files), not about the content that goes in it.
You, as the writer of documentation, are still in charge of its content.
- Put each sentence in its own line.
This makes reviews and suggestions much easier, since GitHub's review system is based on lines.
It also helps identifying long sentences at a glance.
### One sentence per line
- Use the [admonition syntax](#admonitions) for callouts and examples.
Put each sentence in its own line.
This makes reviews and suggestions much easier, since GitHub's review system is based on lines.
It also helps identifying long sentences at a glance.
- Provide at least one example per function, and make examples self-contained.
This is easier to understand for beginners.
It also helps with testing that it actually works especially once we introduce automation.
### Callouts and examples
Example code should be such that it can be passed to `pkgs.callPackage`.
Instead of something like:
Use the [admonition syntax](#admonitions) for callouts and examples.
```nix
pkgs.dockerTools.buildLayeredImage {
name = "hello";
contents = [ pkgs.hello ];
}
```
### Provide self-contained examples
Write something like:
Provide at least one example per function, and make examples self-contained.
This is easier to understand for beginners.
It also helps with testing that it actually works especially once we introduce automation.
```nix
{ dockerTools, hello }:
dockerTools.buildLayeredImage {
name = "hello";
contents = [ hello ];
}
```
Example code should be such that it can be passed to `pkgs.callPackage`.
Instead of something like:
- When showing inputs/outputs of any [REPL](https://en.wikipedia.org/wiki/Read%E2%80%93eval%E2%80%93print_loop), such as a shell or the Nix REPL, use a format as you'd see in the REPL, while trying to visually separate inputs from outputs.
This means that for a shell, you should use a format like the following:
```shell
$ nix-build -A hello '<nixpkgs>' \
--option require-sigs false \
--option trusted-substituters file:///tmp/hello-cache \
--option substituters file:///tmp/hello-cache
/nix/store/zhl06z4lrfrkw5rp0hnjjfrgsclzvxpm-hello-2.12.1
```
Note how the input is preceded by `$` on the first line and indented on subsequent lines, and how the output is provided as you'd see on the shell.
```nix
pkgs.dockerTools.buildLayeredImage {
name = "hello";
contents = [ pkgs.hello ];
}
```
For the Nix REPL, you should use a format like the following:
```shell
nix-repl> builtins.attrNames { a = 1; b = 2; }
[ "a" "b" ]
```
Note how the input is preceded by `nix-repl>` and the output is provided as you'd see on the Nix REPL.
Write something like:
- When documenting functions or anything that has inputs/outputs and example usage, use nested headings to clearly separate inputs, outputs, and examples.
Keep examples as the last nested heading, and link to the examples wherever applicable in the documentation.
```nix
{ dockerTools, hello }:
dockerTools.buildLayeredImage {
name = "hello";
contents = [ hello ];
}
```
The purpose of this convention is to provide a familiar structure for navigating the manual, so any reader can expect to find content related to inputs in an "inputs" heading, examples in an "examples" heading, and so on.
An example:
```
## buildImage
### REPLs
Some explanation about the function here.
Describe a particular scenario, and point to [](#ex-dockerTools-buildImage), which is an example demonstrating it.
When showing inputs/outputs of any [REPL](https://en.wikipedia.org/wiki/Read%E2%80%93eval%E2%80%93print_loop), such as a shell or the Nix REPL, use a format as you'd see in the REPL, while trying to visually separate inputs from outputs.
This means that for a shell, you should use a format like the following:
```shell
$ nix-build -A hello '<nixpkgs>' \
--option require-sigs false \
--option trusted-substituters file:///tmp/hello-cache \
--option substituters file:///tmp/hello-cache
/nix/store/zhl06z4lrfrkw5rp0hnjjfrgsclzvxpm-hello-2.12.1
```
Note how the input is preceded by `$` on the first line and indented on subsequent lines, and how the output is provided as you'd see on the shell.
### Inputs
For the Nix REPL, you should use a format like the following:
```shell
nix-repl> builtins.attrNames { a = 1; b = 2; }
[ "a" "b" ]
```
Note how the input is preceded by `nix-repl>` and the output is provided as you'd see on the Nix REPL.
Documentation for the inputs of `buildImage`.
Perhaps even point to [](#ex-dockerTools-buildImage) again when talking about something specifically linked to it.
### Headings for inputs, outputs and examples
### Passthru outputs
When documenting functions or anything that has inputs/outputs and example usage, use nested headings to clearly separate inputs, outputs, and examples.
Keep examples as the last nested heading, and link to the examples wherever applicable in the documentation.
Documentation for any passthru outputs of `buildImage`.
The purpose of this convention is to provide a familiar structure for navigating the manual, so any reader can expect to find content related to inputs in an "inputs" heading, examples in an "examples" heading, and so on.
An example:
```
## buildImage
### Examples
Some explanation about the function here.
Describe a particular scenario, and point to [](#ex-dockerTools-buildImage), which is an example demonstrating it.
Note that this is the last nested heading in the `buildImage` section.
### Inputs
:::{.example #ex-dockerTools-buildImage}
Documentation for the inputs of `buildImage`.
Perhaps even point to [](#ex-dockerTools-buildImage) again when talking about something specifically linked to it.
# Using `buildImage`
### Passthru outputs
Example of how to use `buildImage` goes here.
Documentation for any passthru outputs of `buildImage`.
:::
```
### Examples
- Use [definition lists](#definition-lists) to document function arguments, and the attributes of such arguments as well as their [types](https://nixos.org/manual/nix/stable/language/values).
For example:
Note that this is the last nested heading in the `buildImage` section.
```markdown
# pkgs.coolFunction {#pkgs.coolFunction}
:::{.example #ex-dockerTools-buildImage}
`pkgs.coolFunction` *`name`* *`config`*
# Using `buildImage`
Description of what `callPackage` does.
Example of how to use `buildImage` goes here.
:::
```
### Function arguments
Use [definition lists](#definition-lists) to document function arguments, and the attributes of such arguments as well as their [types](https://nixos.org/manual/nix/stable/language/values).
For example:
```markdown
# pkgs.coolFunction {#pkgs.coolFunction}
`pkgs.coolFunction` *`name`* *`config`*
Description of what `callPackage` does.
## Inputs {#pkgs-coolFunction-inputs}
## Inputs {#pkgs-coolFunction-inputs}
If something's special about `coolFunction`'s general argument handling, you can say so here.
Otherwise, just describe the single argument or start the arguments' definition list without introduction.
If something's special about `coolFunction`'s general argument handling, you can say so here.
Otherwise, just describe the single argument or start the arguments' definition list without introduction.
*`name`* (String)
*`name`* (String)
: The name of the resulting image.
: The name of the resulting image.
*`config`* (Attribute set)
*`config`* (Attribute set)
: Introduce the parameter. Maybe you have a test to make sure `{ }` is a sensible default; then you can say: these attributes are optional; `{ }` is a valid argument.
: Introduce the parameter. Maybe you have a test to make sure `{ }` is a sensible default; then you can say: these attributes are optional; `{ }` is a valid argument.
`outputHash` (String; _optional_)
`outputHash` (String; _optional_)
: A brief explanation including when and when not to pass this attribute.
: A brief explanation including when and when not to pass this attribute.
: _Default:_ the output path's hash.
```
: _Default:_ the output path's hash.
```
Checklist:
- Start with a synopsis, to show the order of positional arguments.
- Metavariables are in emphasized code spans: ``` *`arg1`* ```. Metavariables are placeholders where users may write arbitrary expressions. This includes positional arguments.
- Attribute names are regular code spans: ``` `attr1` ```. These identifiers can _not_ be picked freely by users, so they are _not_ metavariables.
- _optional_ attributes have a _`Default:`_ if it's easily described as a value.
- _optional_ attributes have a _`Default behavior:`_ if it's not easily described using a value.
- Nix types aren't in code spans, because they are not code
- Nix types are capitalized, to distinguish them from the camelCase Module System types, which _are_ code and behave like functions.
Checklist:
- Start with a synopsis, to show the order of positional arguments.
- Metavariables are in emphasized code spans: ``` *`arg1`* ```. Metavariables are placeholders where users may write arbitrary expressions. This includes positional arguments.
- Attribute names are regular code spans: ``` `attr1` ```. These identifiers can _not_ be picked freely by users, so they are _not_ metavariables.
- _optional_ attributes have a _`Default:`_ if it's easily described as a value.
- _optional_ attributes have a _`Default behavior:`_ if it's not easily described using a value.
- Nix types aren't in code spans, because they are not code
- Nix types are capitalized, to distinguish them from the camelCase Module System types, which _are_ code and behave like functions.
#### Examples