Basically, it doesn't make sense to generate those things every time
you search. That generates a bunch of stuff for the GC to clean up,
when, if the user wanted to do another search, it would just need
to re-do it again.
Every time splice() is called, another temporary object is created.
This version, which uses plain objects as a sort of Hash Bag,
should only produce one temporary object each time it's called.
There is no reason for this function to return an object,
since it is always used for getting at the name anyhow.
It's used in the inner loop for some popular functions,
so we want to avoid allocating in it.
This essentially switches search-index.js from a "array of struct"
to a "struct of array" format, like this:
{
"doc": "Crate documentation",
"t": [ 1, 1, 2, 3, ... ],
"n": [ "Something", "SomethingElse", "whatever", "do_stuff", ... ],
"q": [ "a::b", "", "", "", ... ],
"d": [ "A Struct That Does Something", "Another Struct", "a function", "another function", ... ],
"i": [ 0, 0, 1, 1, ... ],
"f": [ null, null, [], [], ... ],
"p": ...,
"a": ...
}
So `{ty: 1, name: "Something", path: "a::b", desc: "A Struct That Does Something", parent_idx: 0, search_type: null}` is the first item.
This makes the uncompressed version smaller, but it really shows on the
compressed version:
notriddle:rust$ wc -c new-search-index1.52.0.js
2622427 new-search-index1.52.0.js
notriddle:rust$ wc -c old-search-index1.52.0.js
2725046 old-search-index1.52.0.js
notriddle:rust$ gzip new-search-index1.52.0.js
notriddle:rust$ gzip old-search-index1.52.0.js
notriddle:rust$ wc -c new-search-index1.52.0.js.gz
239385 new-search-index1.52.0.js.gz
notriddle:rust$ wc -c old-search-index1.52.0.js.gz
296328 old-search-index1.52.0.js.gz
notriddle:rust$
That's a 4% improvement on the uncompressed version (fewer `[]`),
and 20% improvement after gzipping it, thanks to putting like-typed
data next to each other. Any compression algorithm based on a sliding
window will probably show this kind of improvement.
rustdoc: Remove redundant enableSearchInput function
enableSearchInput was called from two places:
- setupSearchLoader
- addSearchOptions, which is itself called from setupSearchLoader only
This commit can safely get rid of the addSearchOptions calls entirely, and since the setupSearchLoader call is immediately preceded by other method calls on search_input, there's no need to check if it's set.
Both versions are about equally readable, but this version avoids scanning
the entire path and building an intermediate array (`split()` in Rust is
a lazy iterator, but not in JavaScript).
enableSearchInput was called from two places:
- setupSearchLoader
- addSearchOptions, which is itself called from setupSearchLoader only
This commit can safely get rid of the addSearchOptions calls entirely,
and since the setupSearchLoader call is immediately preceded by other
method calls on search_input, there's no need to check if it's set.
It looks like `lev_distance` was used in a very old version of the function,
since it was written but never read, and Blame reports that it was added
before the `checkGenerics` function header itself.
`convertHTMLToPlaintext` is was removed by 768d5e9509
Instead of being loaded on every page, the JS search index is now
loaded when either (a) there is a `?search=` param, or (b) the search
input is focused.
This saves both CPU and bandwidth. As of Feb 2021,
https://doc.rust-lang.org/search-index1.50.0.js is 273,838 bytes
gzipped or 2,544,939 bytes uncompressed. Evaluating it takes 445 ms
of CPU time in Chrome 88 on a i7-10710U CPU (out of a total ~2,100
ms page reload).
Generate separate JS file with crate names.
This is much smaller than the full search index, and is used in the "hot
path" to draw the page. In particular it's used to crate the dropdown
for the search bar, and to append a list of crates to the sidebar (on
some pages).
Skip early search that can bypass 500ms timeout.
This was occurring when someone had typed some text during the load of
search-index.js. Their query was usually not ready to execute, and the
search itself is fairly expensive, delaying the overall load, which
delayed the input / keyup events, which delayed eventually executing the
query.
Remove query parameters when skipping search results
Fixes#81330.
This PR changes the following: when pressing ESC and that no other "action" was performed (understand: no closing the search result, or hiding a menu or something along the line), then we discard the URL query parameters (the `?whatever=dsjfs`). What do you think about this change ```@rust-lang/rustdoc``` ?
EDIT: finally we're simply removing the query parameter when we're skipping the search results.
r? ```@Nemo157```
Improve URLs handling
Fixes#81330.
Explanations: before this PR, when emptying the search input, we still had `?search=` in the URL, which wasn't very nice. Now, if the search is empty, we drop the `?search=` part.
Also, I realized while working on this PR that when we clicked on a menu link when we were on the search results, the search parameters would look like: `?search=#the-anchor`, which was super weird. Now, it looks like this: `?search=the-search#the-anchor`.
Also, I didn't use the `Url` very nice API because it's not available in any IE version (sadness...).
cc `````@lzutao`````
r? `````@Nemo157`````
Improve search result tab handling
Fixes#80378.
If the current search result tab is empty, it picks the first non-empty one. If all are empty, the current one doesn't change. It can be tested with "-> string" (where only the "returned elements" tab is not empty).
r? `@jyn514`
More js cleanup
Part of #79052 (Same kind as #80515).
This one is about some small fixes:
* Replacing some loops with `onEachLazy`.
* Removing unused function arguments.
* Turn `buildHelperPopup` into a variable so it can be "replaced" once the function has been called once so it's not called again.
r? `@jyn514`
Remove tab-lock and replace it with ctrl+up/down arrows to switch between search result tabs
Fixes https://github.com/rust-lang/rust/issues/65212
What took the longest time was to update the help popup in the end.
r? `@Manishearth`
Previously Markdown documentation was not rendered to HTML for search results,
which led to the output not being very readable, particularly for inline code.
This PR fixes that by rendering Markdown to HTML with the help of pulldown-cmark
(the library rustdoc uses to parse Markdown for the main text of documentation).
However, the text for the title attribute (the text shown when you hover over an
element) still uses the plain-text rendering since it is displayed in browsers
as plain-text.
Only these styles will be rendered; everything else is stripped away:
* *italics*
* **bold**
* `inline code`
Make keyboard interactions in the settings menu more pleasant
#78868 improved the keyboard interactions with the settings page. This PR goes a bit further by allowing more than just "space" to toggle the checkboxes.
r? `@jyn514`
Currently, storage.js and main.js have many open-coded calls to
getCurrentValue for "rustdoc-" values, but these are settings and
should be handled by getSettingValue.
So make getSettingValue part of storage.js (where everyone can call
it) and use it everywhere.
No functional change yet. We are going to make getSettingValue do
something more sophisticated in a moment.
Signed-off-by: Ian Jackson <ijackson@chiark.greenend.org.uk>
