src: general C++ cleanup in node_url.cc #19598
Closed
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
- Merge `domain` and `opaque` storage in URL parser: This just simplifies the code a bit, having multiple fields in an union with the same type is usually just overhead. - Add move variant of `URLHost::ToString()`: This helps avoid unnecessary string copy operations, especially since we control the lifetime of `URLHost` objects pretty well. - Use const refs in node_url.cc where appropriate - Remove or reduce overly generous `.reserve()` calls: These would otherwise keep a lot of unused memory lying around. - Return return values instead of unnecessary pointer arguments - Use more common/expressive variable names - Avoid macro use, reduce number of unnecessary JS strings: There’s no reason for `GET`, `GET_AND_SET` and `UTF8STRING` to be macros. Also, `GET` would previously create a JS string instance for each single call, even though the strings it was called with were compile-time constants. - Avoid unnecessary JS casts when the type of a value is known - Avoid (commonly unnecessary) copy for whitespace stripping
nodejs-github-bot
added
c++
TimothyGu
approved these changes
Mar 25, 2018
| @@ -930,7 +925,7 @@ void URLHost::ParseIPv4Host(const char* input, size_t length, bool* is_ipv4) { | |||
| void URLHost::ParseOpaqueHost(const char* input, size_t length) { | |||
| CHECK_EQ(type_, HostType::H_FAILED); | |||
| std::string output; | |||
| output.reserve(length * 3); | |||
| output.reserve(length); | |||
There was a problem hiding this comment.
@TimothyGu i’m not sure whether the spec definition says that this must be ASCII even before percent encoding, but either way, I would personally expect it to be ASCII most or all of the time?
There was a problem hiding this comment.
As long as the tests still pass. :)
src/node_url.cc
Outdated
| bool uflag = false; | ||
| bool atflag = false; // Set when @ has been seen. | ||
| bool square_bracket_flag = false; // Set inside of [...] | ||
| bool inside_password_flag = false; // Set after a : after an username. |
There was a problem hiding this comment.
As a data point, the URL Standard calls it passwordTokenSeenFlag, but seems like you figured out what it meant on yourself :)
There was a problem hiding this comment.
@TimothyGu Idk, do you want me to change it? I think both of these are okay names?
There was a problem hiding this comment.
I'd slightly prefer the name in the spec for better correspondence for future readers.
Yes, and a pretty good one :) But I wanted to wait for the benchmark CI before I started putting numbers on this… ;) |
addaleax
added
whatwg-url
|
Benchmark CI finished. The results are as follows (omitting some files with no or statistically insignificant results): That's some quite substantial speedups!
The only performance decreases look fluky. |
|
Comparison for url/legacy-vs-whatwg-url-parse.js: Before:
After:
|
jasnell
approved these changes
Mar 25, 2018
danbev
approved these changes
Mar 26, 2018
addaleax
added
the
author ready
|
Landed in ae70e2b |
addaleax
removed
the
author ready
addaleax
added a commit
that referenced
this pull request
Mar 30, 2018
- Merge `domain` and `opaque` storage in URL parser: This just simplifies the code a bit, having multiple fields in an union with the same type is usually just overhead. - Add move variant of `URLHost::ToString()`: This helps avoid unnecessary string copy operations, especially since we control the lifetime of `URLHost` objects pretty well. - Use const refs in node_url.cc where appropriate - Remove or reduce overly generous `.reserve()` calls: These would otherwise keep a lot of unused memory lying around. - Return return values instead of unnecessary pointer arguments - Use more common/expressive variable names - Avoid macro use, reduce number of unnecessary JS strings: There’s no reason for `GET`, `GET_AND_SET` and `UTF8STRING` to be macros. Also, `GET` would previously create a JS string instance for each single call, even though the strings it was called with were compile-time constants. - Avoid unnecessary JS casts when the type of a value is known - Avoid (commonly unnecessary) copy for whitespace stripping PR-URL: #19598 Reviewed-By: Tiancheng "Timothy" Gu <[email protected]> Reviewed-By: James M Snell <[email protected]> Reviewed-By: Daniel Bevenius <[email protected]>
targos
pushed a commit
that referenced
this pull request
Apr 2, 2018
- Merge `domain` and `opaque` storage in URL parser: This just simplifies the code a bit, having multiple fields in an union with the same type is usually just overhead. - Add move variant of `URLHost::ToString()`: This helps avoid unnecessary string copy operations, especially since we control the lifetime of `URLHost` objects pretty well. - Use const refs in node_url.cc where appropriate - Remove or reduce overly generous `.reserve()` calls: These would otherwise keep a lot of unused memory lying around. - Return return values instead of unnecessary pointer arguments - Use more common/expressive variable names - Avoid macro use, reduce number of unnecessary JS strings: There’s no reason for `GET`, `GET_AND_SET` and `UTF8STRING` to be macros. Also, `GET` would previously create a JS string instance for each single call, even though the strings it was called with were compile-time constants. - Avoid unnecessary JS casts when the type of a value is known - Avoid (commonly unnecessary) copy for whitespace stripping PR-URL: #19598 Reviewed-By: Tiancheng "Timothy" Gu <[email protected]> Reviewed-By: James M Snell <[email protected]> Reviewed-By: Daniel Bevenius <[email protected]>
Merged
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Merge
domainandopaquestorage in URL parser:This just simplifies the code a bit, having multiple fields
in an union with the same type is usually just overhead.
Add move variant of
URLHost::ToString():This helps avoid unnecessary string copy operations, especially
since we control the lifetime of
URLHostobjects pretty well.Use const refs in node_url.cc where appropriate
Remove or reduce overly generous
.reserve()calls:These would otherwise keep a lot of unused memory lying around.
Return return values instead of unnecessary pointer arguments
Use more common/expressive variable names
Avoid macro use, reduce number of unnecessary JS strings:
There’s no reason for
GET,GET_AND_SETandUTF8STRINGto bemacros. Also,
GETwould previously create a JS string instancefor each single call, even though the strings it was called
with were compile-time constants.
Avoid unnecessary JS casts when the type of a value is known
Avoid (commonly unnecessary) copy for whitespace stripping
Checklist
make -j4 test(UNIX), orvcbuild test(Windows) passes