Rollup merge of rust-lang#107284 - notriddle:notriddle/plus, r=jsha

rustdoc: use smarter encoding for playground URL

The old way would compress okay with DEFLATE, but this version makes uncompressed docs smaller, which matters for memory usage and stuff like `cargo doc`.

Try it out: <https://play.rust-lang.org/?code=fn+main()+{%0Alet+mut+v+=+Vec::new();%0Av.push(1+/+1);%0Aprintln!(%22{}%22,+v[0]);%0A}>

In local testing, this change shrinks sample pages by anywhere between 5.0% and 0.044%

    $ du -b after.dir/std/vec/struct.Vec.html before.dir/std/vec/struct.Vec.html
    753023  after.dir/std/vec/struct.Vec.html
    781842  before.dir/std/vec/struct.Vec.html

100*((753023-781842)/781842)=-3.7

    $ du -b after.dir/std/num/struct.Wrapping.html before.dir/std/num/struct.Wrapping.html
    3189989 after.dir/std/num/struct.Wrapping.html
    3204351 before.dir/std/num/struct.Wrapping.html

100*((3189989-3204351)/3204351)=-0.044

    $ du -b after.dir/std/keyword.match.html before.dir/std/keyword.match.html
    8067    after.dir/std/keyword.match.html
    8495    before.dir/std/keyword.match.html

100*((8067-8495)/8495)=-5.0

Gzipped tarball sizes seem shrunk, but not by much.

    du -s before.tar.gz after.tar.gz
    69600   before.tar.gz
    69492   after.tar.gz

100*((69492-69600)/69600)=-0.16

Author: matthiaskrgr

src/librustdoc/html/markdown.rs

@@ -296,25 +296,59 @@ impl<'a, I: Iterator<Item = Event<'a>>> Iterator for CodeBlocks<'_, 'a, I> {
             let channel = if test.contains("#![feature(") { "&amp;version=nightly" } else { "" };
 
             // These characters don't need to be escaped in a URI.
-            // FIXME: use a library function for percent encoding.
+            // See https://url.spec.whatwg.org/#query-percent-encode-set
+            // and https://url.spec.whatwg.org/#urlencoded-parsing
             fn dont_escape(c: u8) -> bool {
                 (b'a' <= c && c <= b'z')
                     || (b'A' <= c && c <= b'Z')
                     || (b'0' <= c && c <= b'9')
                     || c == b'-'
                     || c == b'_'
                     || c == b'.'
+                    || c == b','
                     || c == b'~'
                     || c == b'!'
                     || c == b'\''
                     || c == b'('
                     || c == b')'
+                    || c == b'['
+                    || c == b']'
+                    || c == b'{'
+                    || c == b'}'
                     || c == b'*'
+                    || c == b'/'
+                    || c == b'|'
+                    || c == b'^'
+                    || c == b'\\'
+                    || c == b';'
+                    || c == b':'
+                    || c == b'?'
+                    // While these would be encoded if we used the URL
+                    // serializer algorithm, they don't actually need to be
+                    // encoded for the [query mode parser] to give the correct
+                    // result. The HTML [attribute parser] does the right
+                    // thing with them as well, as long as the attribute is
+                    // quoted (which it is).
+                    //
+                    // [query mode parser]: https://url.spec.whatwg.org/#query-state
+                    // [attribute parser]: https://html.spec.whatwg.org/#attribute-value-(double-quoted)-state
+                    || c == b'<'
+                    || c == b'>'
+                    // As described in urlencoded-parsing, the
+                    // first `=` is the one that separates key from
+                    // value. Following `=`s are part of the value.
+                    || c == b'='
             }
             let mut test_escaped = String::new();
             for b in test.bytes() {
                 if dont_escape(b) {
                     test_escaped.push(char::from(b));
+                } else if b == b' ' {
+                    // URL queries are decoded with + replaced with SP
+                    test_escaped.push('+');
+                } else if b == b'%' {
+                    test_escaped.push('%');
+                    test_escaped.push('%');
                 } else {
                     write!(test_escaped, "%{:02X}", b).unwrap();
                 }

tests/rustdoc/playground-arg.rs

@@ -10,4 +10,4 @@
 pub fn dummy() {}
 
 // ensure that `extern crate foo;` was inserted into code snips automatically:
-// @matches foo/index.html '//a[@class="test-arrow"][@href="https://example.com/?code=%23!%5Ballow(unused)%5D%0Aextern%20crate%20r%23foo%3B%0Afn%20main()%20%7B%0Ause%20foo%3A%3Adummy%3B%0Adummy()%3B%0A%7D&edition=2015"]' "Run"
+// @matches foo/index.html '//a[@class="test-arrow"][@href="https://example.com/?code=%23![allow(unused)]%0Aextern+crate+r%23foo;%0Afn+main()+{%0Ause+foo::dummy;%0Adummy();%0A}&edition=2015"]' "Run"

tests/rustdoc/playground.rs

@@ -22,6 +22,6 @@
 //! }
 //! ```
 
-// @matches foo/index.html '//a[@class="test-arrow"][@href="https://www.example.com/?code=%23!%5Ballow(unused)%5D%0Afn%20main()%20%7B%0A%20%20%20%20println!(%22Hello%2C%20world!%22)%3B%0A%7D&edition=2015"]' "Run"
-// @matches foo/index.html '//a[@class="test-arrow"][@href="https://www.example.com/?code=%23!%5Ballow(unused)%5D%0Afn%20main()%20%7B%0Aprintln!(%22Hello%2C%20world!%22)%3B%0A%7D&edition=2015"]' "Run"
-// @matches foo/index.html '//a[@class="test-arrow"][@href="https://www.example.com/?code=%23!%5Ballow(unused)%5D%0A%23!%5Bfeature(something)%5D%0A%0Afn%20main()%20%7B%0A%20%20%20%20println!(%22Hello%2C%20world!%22)%3B%0A%7D&version=nightly&edition=2015"]' "Run"
+// @matches foo/index.html '//a[@class="test-arrow"][@href="https://www.example.com/?code=%23![allow(unused)]%0Afn+main()+{%0Aprintln!(%22Hello,+world!%22);%0A}&edition=2015"]' "Run"
+// @matches foo/index.html '//a[@class="test-arrow"][@href="https://www.example.com/?code=%23![allow(unused)]%0Afn+main()+{%0A++++println!(%22Hello,+world!%22);%0A}&edition=2015"]' "Run"
+// @matches foo/index.html '//a[@class="test-arrow"][@href="https://www.example.com/?code=%23![allow(unused)]%0A%23![feature(something)]%0A%0Afn+main()+{%0A++++println!(%22Hello,+world!%22);%0A}&version=nightly&edition=2015"]' "Run"