Add the rustdoc readme #86

mark-i-m · 2018-03-13T00:52:15Z

After this merges, I hope to remove the README from librustdoc so that we can have everything in one place.

mark-i-m · 2018-03-13T00:53:30Z

I should note that (except for the first sentence) this is a straight copy/paste of what @QuietMisdreavus wrote up.

mark-i-m · 2018-03-13T00:54:51Z

cc #2

QuietMisdreavus

A few notes from when i originally wrote this, and questions about integrating it here. I'm not familiar with any processes y'all may have set up, so i'll just leave these remarks as "if you want to change this, here's more information".

Thanks for porting this! I'm glad that we can get rustdoc into the guide here.

QuietMisdreavus · 2018-03-13T14:26:21Z

src/rustdoc.md

+Naturally, there's more than just this, and those descriptions simplify out lots of details, but
+that's the high-level overview.
+
+(Side note: this is a library crate! The `rustdoc` binary is crated using the project in


Since the readme is no longer right next to the code, it would probably be better to rephrase this line to "Side note: librustdoc is a library crate!" to avoid confusion.

QuietMisdreavus · 2018-03-13T15:43:38Z

src/rustdoc.md

+[under way]: https://github.com/steveklabnik/rustdoc
+
+Rustdoc is implemented entirely within the crate `librustdoc`. After partially compiling a crate to
+get its AST (technically the HIR map) from rustc, librustdoc performs two major steps past that to


partially compiling a crate to get its AST (technically the HIR map)

Is this a valid thing to say? The complete story of what librustdoc does is that it runs "phase 3" of the compilation process to get a TyCtxt and query the compiler about type information that it can't get from the source code. But saying that it compiles the crate to its HIR felt like a simpler/smoother thing to say in this context.

What about saying something along the lines of "it runs the compiler up to the point where we have an internal representation of a crate (HIR) and the ability to query the type system (via TyCtxt)"?

I'm not sure referring to the TyCtxt as "the type system" is actually correct, but it helps give you the general idea.

QuietMisdreavus · 2018-03-13T15:45:09Z

src/rustdoc.md

+
+### From soup to nuts
+
+(alternate title: ["An unbroken thread that stretches from those first `Cell`s to us"][video])


When i first wrote this, i waffled about whether or not to include this reference. My eventual reasoning was that READMEs weren't as formal a context as proper crate documentation, so it was okay to sneak this pun in. Now that the document's being moved here, is this worth keeping?

I like the light-hearted title 😁

It's kinda hard to put a simple name to this section, so I'm not sure what I'd replace it with if we were to change it. We're essentially trying to exfiltrate information from the guts of the compiler without breaking things due to lifetimes and the various other restrictions.

To be fair, the entire document is full of my usual style of "vague headings that only kinda relate to their section via some extended metaphor" blog-post writing. The specific link to the video is because of the Cell pun - DocContext and SharedContext and Cache are full of RefCells because of how they're passed around. "From soup to nuts" is a variation of the idea that the thing being discussed stretches across the whole documentation process, and won out because it fits the rhythm i established with "From crate to clean" and "From clean to crate". Whether or not to keep the video link is up to the accumulated style of the rest of guide, which is an editorial decision i'm not prepared to make. >_>

I personally like the titles :)

We can change the later if others want a more formal tone...

QuietMisdreavus · 2018-03-13T15:47:29Z

src/rustdoc.md

+"`find_testable_code`" in `html/markdown.rs`, it builds up a collection of tests to run before
+handing them off to the libtest test runner. One notable location in `test.rs` is the function
+`make_test`, which is where hand-written doctests get transformed into something that can be
+executed.


I have a longer writeup about make_test if y'all feel it's worth linking here.

Its probably a good idea to mention it. I just skimmed through your article and learner a lot about how rustdoc generates tests. It's really quite smart.

Also, totally unrelated, but that gives me an idea for how to build a simple REPL for rust :P

Michael-F-Bryan · 2018-03-14T14:45:54Z

src/rustdoc.md

+Before moving on to the next major step, a few important "passes" occur over the documentation.
+These do things like combine the separate "attributes" into a single string and strip leading
+whitespace to make the document easier on the markdown parser, or drop items that are not public or
+deliberately hidden with `#[doc(hidden)]`. These are all implemented in the `passes/` directory, one


Is this a duplicate of the sassy little "I actually liked a little earlier..." paragraph?

Both that paragraph and this one talk about walking the crate and removing/altering things according to the [doc(...)] attributes.

The key difference between the "I lied a littler earlier" and the "Hot potato" paragraphs is that of timing: The visit_ast.rs processing happens before crate cleaning, but the passes run after it (but before page rendering).

Michael-F-Bryan · 2018-03-14T15:00:00Z

src/rustdoc.md

+
+### From soup to nuts
+
+(alternate title: ["An unbroken thread that stretches from those first `Cell`s to us"][video])


I like the light-hearted title 😁

It's kinda hard to put a simple name to this section, so I'm not sure what I'd replace it with if we were to change it. We're essentially trying to exfiltrate information from the guts of the compiler without breaking things due to lifetimes and the various other restrictions.

Michael-F-Bryan · 2018-03-14T15:00:48Z

src/rustdoc.md

+
+[under way]: https://github.com/steveklabnik/rustdoc
+
+Rustdoc is implemented entirely within the crate `librustdoc`. After partially compiling a crate to


Link to the crate root on GitHub?

Michael-F-Bryan · 2018-03-14T15:02:04Z

src/rustdoc.md

+
+## From crate to clean
+
+In `core.rs` are two central items: the `DocContext` struct, and the `run_core` function. The latter


It may be worth adding DocContext to the code index at the end of the guide.

Michael-F-Bryan · 2018-03-14T15:02:45Z

src/rustdoc.md

+`#[doc=""]` attributes into a separate field of the Attributes struct, present on anything that gets
+hand-written documentation. This makes it easier to collect this documentation later in the process.
+
+The primary output of this process is a clean::Crate with a tree of Items which describe the


Should this be clean::Crate?

No, I believe it is this one: https://github.com/rust-lang/rust/blob/master/src/librustdoc/clean/mod.rs#L132

Oh no. I meant there should be back ticks around the word so it gets styles as code.

Yes, this should get backticks since the rest of the document uses those around other type names.

Michael-F-Bryan · 2018-03-14T15:04:43Z

src/rustdoc.md

+### Hot potato
+
+Before moving on to the next major step, a few important "passes" occur over the documentation.
+These do things like combine the separate "attributes" into a single string and strip leading


If we're listing the various rustdoc passes, would it be worth using bullet points? E.g.

collate and clean doc comments

remove [doc(hidden)] items

...

Here's the current list of passes for rustdoc, as given by a kinda-recent nightly:

$ rustdoc +nightly --passes list WARNING: the 'passes' flag is considered deprecated WARNING: please see https://github.com/rust-lang/rust/issues/44136 Available passes for running rustdoc: strip-hidden - strips all doc(hidden) items from the output unindent-comments - removes excess indentation on comments in order for markdown to like it collapse-docs - concatenates all document attributes into one document attribute strip-private - strips all private items from a crate which cannot be seen externally, implies strip-priv-imports strip-priv-imports - strips all private import statements (`use`, `extern crate`) from a crate propagate-doc-cfg - propagates `#[doc(cfg(...))]` to child items Default passes for rustdoc: strip-hidden strip-private collapse-docs unindent-comments propagate-doc-cfg

collapse-docs is necessary because each line of a doc comment is given as a separate doc attribute, and this will combine them into a single string with line breaks between each attribute.

unindent-comments is necessary because the convention for writing documentation is to provide a space between the /// or //! marker and the text, and stripping that leading space will make the text easier to parse by the Markdown parser. (In my experience it's less necessary now that we have a Commonmark-compliant parser, since it doesn't have a problem with headers that have a space before the ## that marks the heading.)

strip-priv-imports is necessary because rustdoc will handle public imports by either inlining the item's documentation to the module or creating a "Reexports" section with the import in it. The pass ensures that all of these imports are actually relevant to documentation.

strip-hidden and strip-private also remove items that are not relevant for public documentation.

Passing --document-private-items disables all the passes and only runs collapse-docs and unindent-comments, which means that everything in the crate gets documented, not just the public items.

I added this list (verbatim from your comment) to the chapter, but I think we should avoid documenting some of these hidden cmd line flags since they seem unstable...

Michael-F-Bryan · 2018-03-14T15:05:44Z

src/rustdoc.md

+whitespace to make the document easier on the markdown parser, or drop items that are not public or
+deliberately hidden with `#[doc(hidden)]`. These are all implemented in the `passes/` directory, one
+file per pass. By default, all of these passes are run on a crate, but the ones regarding dropping
+private/hidden items can be bypassed by passing `--document-private-items` to rustdoc.


Ooh, that's cool! I wonder if you can use it from cargo.

Not directly. You'd have to use $RUSTDOCFLAGS or cargo rustdoc to add additional arguments to rustdoc.

Michael-F-Bryan · 2018-03-14T15:06:38Z

src/rustdoc.md

+## From clean to crate
+
+This is where the "second phase" in rustdoc begins. This phase primarily lives in the `html/`
+folder, and it all starts with `run()` in `html/render.rs`. This code is responsible for setting up


Do you reckon this file would be noteworthy enough to warrant a link to its page on GitHub?

Hmm... it would be a bit weird IMHO since there are a lot of other files mentioned, and none of them got special treatment...

mark-i-m · 2018-03-15T03:08:37Z

@Michael-F-Bryan @QuietMisdreavus I think I addressed all of the comments. Let me know if I missed anything.

Michael-F-Bryan · 2018-03-15T08:10:39Z

Looks good to me. @QuietMisdreavus, do you think this is good to merge?

QuietMisdreavus

A couple more comments.

QuietMisdreavus · 2018-03-15T13:59:35Z

src/appendix-code-index.md

@@ -8,6 +8,7 @@ Item            |  Kind    | Short description           | Chapter            |
 ----------------|----------|-----------------------------|--------------------|-------------------
 `CodeMap` | struct | The CodeMap maps the AST nodes to their source code | [The parser] | [src/libsyntax/codemap.rs](https://github.com/rust-lang/rust/blob/master/src/libsyntax/codemap.rs)
 `CompileState` | struct | State that is passed to a callback at each compiler pass | [The Rustc Driver] | [src/librustc_driver/driver.rs](https://github.com/rust-lang/rust/blob/master/src/librustc_driver/driver.rs)
+`DocContext` | struct | A state container used by rustdoc when crawling through a crate to gather its documentation | [Rustdoc] | [src/librustdoc/core.rs](https://github.com/rust-lang/rust/blob/master/src/librustdoc/core.rs)


[Rustdoc] doesn't have a link target.

QuietMisdreavus · 2018-03-15T14:02:12Z

src/rustdoc.md


 (Strictly speaking, you can fine-tune the passes run and even add your own, but [we're trying to
 deprecate that][44136]. If you need finer-grain control over these passes, please let us know!)

 [44136]: https://github.com/rust-lang/rust/issues/44136

+Here is current (as of this writing) list of passes:
+
+- `collapse-docs` is necessary because each line of a doc comment is given as a


The structure of this list a little jarring because it's introduced as "here are the passes" but it's written as "why are they here". Specifically, starting off with "(pass) is necessary..." is awkward because you haven't even described what the pass does yet.

QuietMisdreavus · 2018-03-15T14:03:27Z

src/rustdoc.md

+`#[doc=""]` attributes into a separate field of the Attributes struct, present on anything that gets
+hand-written documentation. This makes it easier to collect this documentation later in the process.
+
+The primary output of this process is a clean::Crate with a tree of Items which describe the


Yes, this should get backticks since the rest of the document uses those around other type names.

mark-i-m · 2018-03-15T14:40:36Z

@QuietMisdreavus I think I have addressed your comments. I added a bit to each bullet to describe what it does first. Let me know what you think :)

QuietMisdreavus · 2018-03-15T14:54:11Z

src/rustdoc.md

  separate doc attribute, and this will combine them into a single string with
  line breaks between each attribute.
- `unindent-comments` is necessary because the convention for writing
+- `unindent-comments` removes excess indentation on comments in order for
+  markdown to like it. This is necessary because the convention for writing
  documentation is to provide a space between the `///` or `//!` marker and the
  text, and stripping that leading space will make the text easier to parse by
  the Markdown parser. (In my experience it's less necessary now that we have a
  Commonmark-compliant parser, since it doesn't have a problem with headers
  that have a space before the `##` that marks the heading.)


I'm not sure this parenthetical is worth keeping? It's definitely not adding as-is, since the rest of the document has so far avoided first-person. Maybe having a historical note that the previous Markdown parser had a problem with spaces before a heading would be useful, but i'm less confident about the remark about it being "less necessary" since i'm not sure that Pulldown is totally okay with everything having an indent on it.

QuietMisdreavus · 2018-03-15T18:11:37Z

Looks good to me! Sorry about the back-and-forth there at the end. This has my 👍 now.

mark-i-m · 2018-03-15T20:00:05Z

Sorry about the back-and-forth there at the end.

Not at all! Thanks for taking the time 😄

…dreavus Move librustdoc readme to rustc guide cc rust-lang/rustc-dev-guide#2 and rust-lang#48478 Don't merge this before rust-lang/rustc-dev-guide#86

Add the rustdoc readme

96fd186

nikomatsakis mentioned this pull request Mar 13, 2018

move over the rustc README #2

Closed

19 tasks

mark-i-m mentioned this pull request Mar 13, 2018

Move librustdoc readme to rustc guide rust-lang/rust#48972

Merged

QuietMisdreavus reviewed Mar 13, 2018

View reviewed changes

Michael-F-Bryan reviewed Mar 14, 2018

View reviewed changes

Address review comments

5f9f748

QuietMisdreavus reviewed Mar 15, 2018

View reviewed changes

address review comments

881e307

QuietMisdreavus reviewed Mar 15, 2018

View reviewed changes

Updated parenthetical

c32bc39

mark-i-m merged commit 81071c6 into rust-lang:master Mar 15, 2018

mark-i-m deleted the rustdoc branch May 23, 2018 16:42


		### From soup to nuts

		(alternate title: ["An unbroken thread that stretches from those first `Cell`s to us"][video])


		[under way]: https://github.com/steveklabnik/rustdoc

		Rustdoc is implemented entirely within the crate `librustdoc`. After partially compiling a crate to


		## From crate to clean

		In `core.rs` are two central items: the `DocContext` struct, and the `run_core` function. The latter

Add the rustdoc readme #86

Add the rustdoc readme #86

Conversation

mark-i-m commented Mar 13, 2018

mark-i-m commented Mar 13, 2018

mark-i-m commented Mar 13, 2018

QuietMisdreavus left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

QuietMisdreavus Mar 14, 2018 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

mark-i-m commented Mar 15, 2018

Michael-F-Bryan commented Mar 15, 2018

QuietMisdreavus left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

mark-i-m commented Mar 15, 2018

Choose a reason for hiding this comment

Choose a reason for hiding this comment

QuietMisdreavus commented Mar 15, 2018

mark-i-m commented Mar 15, 2018

QuietMisdreavus Mar 14, 2018 •

edited

Loading