Files
Umbraco-CMS/tests/Umbraco.Tests.Benchmarks/StringExtensionsBenchmarks.cs
Kenn Jacobsen 3bd66b89e1 Merge branch 'v15/dev' into v16/dev (#18971)
* Only prevent the unpublish or delete of a related item when configured to do so if it is related as a child, not as a parent (#18886)

* Only prevent the unpubkish or delete of a related item when configured to do so if it is related as a child, not as a parent.

* Fixed incorect parameter names.

* Fixed failing integration tests.

* Use using variable instead to reduce nesting

* Applied suggestions from code review.

* Used simple using statement throughout RelationService for consistency.

* Applied XML header comments consistently.

---------

Co-authored-by: mole <nikolajlauridsen@protonmail.ch>

* Feature: highlight invariant doc with variant blocks is unsupported (#18806)

* mark variant blocks in invariant docs as invalid

* implement RTE Blocks

* Fix pagination for users restricted by start nodes (#18907)

* Fix pagination for users restricted by start nodes

* Default implementation to avoid breakage

* Review comments

* Fix failing test

* Add media start node tests

* Fix issue preventing blueprint derived values from being scaffolded (#18917)

* Fix issue preventing blueprint derived values from being scaffolded.

* fix manipulating frooen array

* compare with variantId as well

---------

Co-authored-by: Niels Lyngsø <niels.lyngso@gmail.com>

* ci: add Azure Static Web Apps workflow file
on-behalf-of: @Azure opensource@microsoft.com

* ci: add Azure Static Web Apps workflow file
on-behalf-of: @Azure opensource@microsoft.com

* ci: add Azure Static Web Apps workflow file
on-behalf-of: @Azure opensource@microsoft.com

* Remove admin permission on user configuration, allowing users with user section access only to manaage users and groups. (#18848)

* Tiptap RTE: Style Menu extension kind (#18918)

* Adds 'styleMenu' Tiptap toolbar extension kind

* Adds icons for `<h4>` and `<p>` tags

* Adds commands to HTML Global Attributes extension

for setting the `class` and `id` attributes.

* Renamed "default-tiptap-toolbar-element.api.ts" file

The "element" part was confusing.

* Toolbar Menu: uses correct `item` value

* Cascading Menu: adds localization for the label

* Adds `label` attribute to UUI components

for accessibility.

* Toolbar Menu: uses correct `appearance` value

* Removed unrequired `api` from Style Select

* Destructs the `item.data` object

* Ensure has children reflects only items with folder children when folders only are queried. (#18790)

* Ensure has children reflects only items with folder children when folders only are queried.

* Added supression for change to integration test public code.

---------

Co-authored-by: Migaroez <geusens@gmail.com>

* Only apply validation on content update to variant cultures where the editor has permission for the culture (#18778)

* Only apply validation on content update to variant cultures where the editor has permission for the culture.

* Remove inadvertent comment updates.

* Fixed failing integration test.

* Adds ancestor ID details on document tree and collection responses (#18909)

* Populate ancestor keys on document tree response items.

* Populate ancestor keys on document collection response items.

* Update OpenApi.json

* Use array of objects rather than Ids for the ancestor collection.

* Update OpenApi.json.

* Move publish with descendants to a background task with polling (#18497)

* Use background queue for database cache rebuild and track rebuilding status.

* Updated OpenApi.json and client-side types.

* Updated client to poll for completion of database rebuild.

* Move IBackgroundTaskQueue to core and prepare publish branch to run as background task.

* Endpoints for retrieval of status and result from branch publish operations.

* Poll and retrieve result for publish with descendants.

* Handled issues from testing.

* Rework to single controller for status and result.

* Updated client side sdk.

* OpenApi post dev merge gen

---------

Co-authored-by: Migaroez <geusens@gmail.com>

* Clear roots before rebuilding navigation dictionary (#18766)

* Clear roots before rebuilding navigation dictionary.

* Added tests to verify fix.

* Correct test implementation.

* Convert integration tests with method overloads into test cases.

* Integration test compatibility supressions.

* Fixes save of empty, invariant block list on variant content. (#18932)

* remove unnecessary code (#18927)

* V15/bugfix/fix route issue from 18859 (#18931)

* unique check

* unique for workspace empty path

* more unique routes

* Bump vite from 6.2.3 to 6.2.4 in /src/Umbraco.Web.UI.Client

Bumps [vite](https://github.com/vitejs/vite/tree/HEAD/packages/vite) from 6.2.3 to 6.2.4.
- [Release notes](https://github.com/vitejs/vite/releases)
- [Changelog](https://github.com/vitejs/vite/blob/v6.2.4/packages/vite/CHANGELOG.md)
- [Commits](https://github.com/vitejs/vite/commits/v6.2.4/packages/vite)

---
updated-dependencies:
- dependency-name: vite
  dependency-version: 6.2.4
  dependency-type: direct:development
...

Signed-off-by: dependabot[bot] <support@github.com>

* removes autogenerated workflows

* make getHasUnpersistedChanges public (#18929)

* Added management API endpoint, service and repository for retrieval of references from the recycle bin (#18882)

* Added management API endpoint, service and repository for retrieval of references from the recycle bin.

* Update src/Umbraco.Cms.Api.Management/Controllers/Document/RecycleBin/ReferencedByDocumentRecycleBinController.cs

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* Removed unused code.

---------

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* Updated management API endpoint and model for data type references to align with that used for documents, media etc. (#18905)

* Updated management API endpoint and model for data type references to align with that used for documents, media etc.

* Refactoring.

* Update src/Umbraco.Core/Constants-ReferenceTypes.cs

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* Fixed typos.

* Added id to tracked reference content type response.

* Updated OpenApi.json.

* Added missing updates.

* Renamed model and constants from code review feedback.

* Fix typo

* Fix multiple enumeration

---------

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Co-authored-by: mole <nikolajlauridsen@protonmail.ch>

* Skip lock tests

* Look-up redirect in content finder for multi-lingual sites using path and legacy route prefixed with the integer ID of the node with domains defined (#18763)

* Look-up redirect in content finder for multi-lingual sites using path and legacy route prefixed with the integer ID of the node with domains defined.

* Added tests to verify functionality.

* Added reference to previous PR.

* Referenced second PR.

* Assemble URLs for all cultures, not just the default.

* Revert previous update.

* Display an original URL if we have one.

* Bump vite from 6.2.4 to 6.2.5 in /src/Umbraco.Web.UI.Client

Bumps [vite](https://github.com/vitejs/vite/tree/HEAD/packages/vite) from 6.2.4 to 6.2.5.
- [Release notes](https://github.com/vitejs/vite/releases)
- [Changelog](https://github.com/vitejs/vite/blob/v6.2.5/packages/vite/CHANGELOG.md)
- [Commits](https://github.com/vitejs/vite/commits/v6.2.5/packages/vite)

---
updated-dependencies:
- dependency-name: vite
  dependency-version: 6.2.5
  dependency-type: direct:development
...

Signed-off-by: dependabot[bot] <support@github.com>

* Add raw value validation to multiple text strings property editor (#18936)

* Add raw value validation to multiple text strings property editor

* Added additional assert on unit test and comment on validation logic.

* Don't remove items to obtain a valid value

---------

Co-authored-by: Andy Butland <abutland73@gmail.com>

* Integration tests for content publishing with ancestor unpublished (#18941)

* Resolved warnings in test class.

* Refactor regions into partial classes.

* Aligned test names.

* Variable name refactoring.

* Added tests for unpublished paths.

* Adjust tests to verify current behaviour.

* Cleaned up project file.

* fix circular icon import (#18952)

* remove segment toggle for elements (#18949)

* Fix modal route registration circular import (#18953)

* fix modal route registration circular import

* Update modal-route-registration.controller.ts

* V15/fix/18595 (#18925)

* fix for #18595

* updates the en.ts

* Avoid unneeded Dictionary operations (#18890)

* Avoid some heap allocations

* Remove unneeded double seek

* Avoid allocating new empty arrays, reuse existing empty array

* Avoid allocating strings for parsing comma separated int values (#18199)

* Data type References UI: Workspace + Delete (#18914)

* Updated management API endpoint and model for data type references to align with that used for documents, media etc.

* Refactoring.

* Update src/Umbraco.Core/Constants-ReferenceTypes.cs

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* Fixed typos.

* generate server models

* add extension slot

* register data type reference info app

* add reference data mappers

* Added id to tracked reference content type response.

* Updated OpenApi.json.

* Added missing updates.

* generate new models

* update models

* register ref item

* remove debugger

* render types

* register member type property type ref

* register media type property type ref

* Renamed model and constants from code review feedback.

* register reference workspace info app kind

* use kind for document references

* use kind for media references

* use kind for member references

* use deleteWithRelation kind when deleting data types

* fix manifest types

* fix types

* Update types.gen.ts

* update code to fit new server models

---------

Co-authored-by: Andy Butland <abutland73@gmail.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* Feature: discard changes for block workspace (#18930)

* make getHasUnpersistedChanges public

* Discard changes impl for Block Workspace

* fix 18367 (#18956)

* Merge commit from fork

* Prevent path traveral vulnerability with upload of temporary files.

* Used BadRequest instead of NotFound for invalid file name response.

* V15 QA Fixing the failing media acceptance tests (#18881)

* Fixed the function name due to test helper changes

* Updated assertion steps due to UI changes

* Added more waits

* Bumped version

* Increase timeout

* Reverted

---------

Co-authored-by: Andreas Zerbst <73799582+andr317c@users.noreply.github.com>

* V15 QA added clipboard test for not being able to copy to root when block is not allowed at root (#18937)

* Added clipboard test

* Bumped version

* Updated to use the name

* Run all tests on the pipeline

* Reverted command

* build: adjusts circular ref number to 4

---------

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: Andy Butland <abutland73@gmail.com>
Co-authored-by: mole <nikolajlauridsen@protonmail.ch>
Co-authored-by: Niels Lyngsø <nsl@umbraco.dk>
Co-authored-by: Niels Lyngsø <niels.lyngso@gmail.com>
Co-authored-by: Jacob Overgaard <752371+iOvergaard@users.noreply.github.com>
Co-authored-by: Lee Kelleher <leekelleher@users.noreply.github.com>
Co-authored-by: Migaroez <geusens@gmail.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Co-authored-by: Nikolaj Geisle <70372949+Zeegaan@users.noreply.github.com>
Co-authored-by: Mads Rasmussen <madsr@hey.com>
Co-authored-by: Jacob Welander Jensen <64834767+Welander1994@users.noreply.github.com>
Co-authored-by: Henrik <hg@impact.dk>
Co-authored-by: Sebastiaan Janssen <sebastiaan@umbraco.com>
Co-authored-by: Nhu Dinh <150406148+nhudinh0309@users.noreply.github.com>
Co-authored-by: Andreas Zerbst <73799582+andr317c@users.noreply.github.com>
2025-04-09 09:58:01 +02:00

258 lines
11 KiB
C#

using System.Globalization;
using System.Runtime.InteropServices;
using BenchmarkDotNet.Attributes;
using Umbraco.Cms.Core;
namespace Umbraco.Tests.Benchmarks;
[MemoryDiagnoser]
public class StringExtensionsBenchmarks
{
private static readonly Random _seededRandom = new(60);
private const int Size = 100;
private static readonly string[] _stringsWithCommaSeparatedNumbers = new string[Size];
static StringExtensionsBenchmarks()
{
for (var i = 0; i < Size; i++)
{
int countOfNumbers = _seededRandom.Next(2, 10); // guess on path lengths in normal use
int[] randomIds = new int[countOfNumbers];
for (var i1 = 0; i1 < countOfNumbers; i1++)
{
randomIds[i1] = _seededRandom.Next(-1, int.MaxValue);
}
_stringsWithCommaSeparatedNumbers[i] = string.Join(',', randomIds);
}
}
/// <summary>
/// Ye olden way of doing it (before 20250201 https://github.com/umbraco/Umbraco-CMS/pull/18048)
/// </summary>
/// <returns>A number so the compiler/runtime doesn't optimize it away.</returns>
[Benchmark]
public int Linq()
{
var totalNumberOfIds = 0; // a number to operate on so it is not optimized away
foreach (string? stringWithCommaSeparatedNumbers in _stringsWithCommaSeparatedNumbers)
{
totalNumberOfIds += Linq(stringWithCommaSeparatedNumbers).Length;
}
return totalNumberOfIds;
}
private static int[] Linq(string path)
{
int[]? nodeIds = path.Split(Constants.CharArrays.Comma, StringSplitOptions.RemoveEmptyEntries)
.Select(x =>
int.TryParse(x, NumberStyles.Integer, CultureInfo.InvariantCulture, out var output)
? Attempt<int>.Succeed(output)
: Attempt<int>.Fail())
.Where(x => x.Success)
.Select(x => x.Result)
.Reverse()
.ToArray();
return nodeIds;
}
/// <summary>
/// Here we are allocating strings to the separated values,
/// BUT we know the count of numbers, so we can allocate the exact size of list we need
/// </summary>
/// <returns>A number so the compiler/runtime doesn't optimize it away.</returns>
[Benchmark]
public int SplitToHeapStrings()
{
var totalNumberOfIds = 0; // a number to operate on so it is not optimized away
foreach (string stringWithCommaSeparatedNumbers in _stringsWithCommaSeparatedNumbers)
{
totalNumberOfIds += SplitToHeapStrings(stringWithCommaSeparatedNumbers).Length;
}
return totalNumberOfIds;
}
private static int[] SplitToHeapStrings(string path)
{
string[] pathSegments = path.Split(Constants.CharArrays.Comma, StringSplitOptions.RemoveEmptyEntries);
List<int> nodeIds = new(pathSegments.Length); // here we know how large the resulting list should at least be
for (int i = pathSegments.Length - 1; i >= 0; i--)
{
if (int.TryParse(pathSegments[i], NumberStyles.Integer, CultureInfo.InvariantCulture, out int pathSegment))
{
nodeIds.Add(pathSegment);
}
}
return nodeIds.ToArray(); // allocates a new array
}
/// <summary>
/// Here we avoid allocating strings to the separated values,
/// BUT we do not know the count of numbers, so we might end up resizing the list we add numbers to it
/// </summary>
/// <returns>A number so the compiler/runtime doesn't optimize it away.</returns>
[Benchmark]
public int SplitToStackSpansWithoutEmptyCheckReversingListAsSpan()
{
var totalNumberOfIds = 0; // a number to operate on so it is not optimized away
foreach (string stringWithCommaSeparatedNumbers in _stringsWithCommaSeparatedNumbers)
{
totalNumberOfIds += SplitToStackSpansWithoutEmptyCheckReversingListAsSpan(stringWithCommaSeparatedNumbers).Length;
}
return totalNumberOfIds;
}
private static int[] SplitToStackSpansWithoutEmptyCheckReversingListAsSpan(string path)
{
ReadOnlySpan<char> pathSpan = path.AsSpan();
MemoryExtensions.SpanSplitEnumerator<char> pathSegments = pathSpan.Split(Constants.CharArrays.Comma);
// Here we do NOT know how large the resulting list should at least be
// Default empty List<> internal array capacity on add is currently 4
// If the count of numbers are less than 4, we overallocate a little
// If the count of numbers are more than 4, the list will be resized, resulting in a copy from initial array to new double size array
// If the count of numbers are more than 8, another new array is allocated and copied to
List<int> nodeIds = [];
foreach (Range rangeOfPathSegment in pathSegments)
{
// this is only a span of the string, a string is not allocated on the heap
ReadOnlySpan<char> pathSegmentSpan = pathSpan[rangeOfPathSegment];
if (int.TryParse(pathSegmentSpan, NumberStyles.Integer, CultureInfo.InvariantCulture, out int pathSegment))
{
nodeIds.Add(pathSegment);
}
}
Span<int> nodeIdsSpan = CollectionsMarshal.AsSpan(nodeIds);
var result = new int[nodeIdsSpan.Length];
var resultIndex = 0;
for (int i = nodeIdsSpan.Length - 1; i >= 0; i--)
{
result[resultIndex++] = nodeIdsSpan[i];
}
return result;
}
/// <summary>
/// Here we avoid allocating strings to the separated values,
/// BUT we do not know the count of numbers, so we might end up resizing the list we add numbers to it
/// </summary>
/// <returns>A number so the compiler/runtime doesn't optimize it away.</returns>
[Benchmark]
public int SplitToStackSpansWithoutEmptyCheck()
{
var totalNumberOfIds = 0; // a number to operate on so it is not optimized away
foreach (string stringWithCommaSeparatedNumbers in _stringsWithCommaSeparatedNumbers)
{
totalNumberOfIds += SplitToStackSpansWithoutEmptyCheck(stringWithCommaSeparatedNumbers).Length;
}
return totalNumberOfIds;
}
private static int[] SplitToStackSpansWithoutEmptyCheck(string path)
{
ReadOnlySpan<char> pathSpan = path.AsSpan();
MemoryExtensions.SpanSplitEnumerator<char> pathSegments = pathSpan.Split(Constants.CharArrays.Comma);
// Here we do NOT know how large the resulting list should at least be
// Default empty List<> internal array capacity on add is currently 4
// If the count of numbers are less than 4, we overallocate a little
// If the count of numbers are more than 4, the list will be resized, resulting in a copy from initial array to new double size array
// If the count of numbers are more than 8, another new array is allocated and copied to
List<int> nodeIds = [];
foreach (Range rangeOfPathSegment in pathSegments)
{
// this is only a span of the string, a string is not allocated on the heap
ReadOnlySpan<char> pathSegmentSpan = pathSpan[rangeOfPathSegment];
if (int.TryParse(pathSegmentSpan, NumberStyles.Integer, CultureInfo.InvariantCulture, out int pathSegment))
{
nodeIds.Add(pathSegment);
}
}
var result = new int[nodeIds.Count];
var resultIndex = 0;
for (int i = nodeIds.Count - 1; i >= 0; i--)
{
result[resultIndex++] = nodeIds[i];
}
return result;
}
/// <summary>
/// Here we avoid allocating strings to the separated values,
/// BUT we do not know the count of numbers, so we might end up resizing the list we add numbers to it
/// </summary>
/// <remarks>Here with an empty check, unlikely in umbraco use case.</remarks>
/// <returns>A number so the compiler/runtime doesn't optimize it away.</returns>
[Benchmark]
public int SplitToStackSpansWithEmptyCheck()
{
var totalNumberOfIds = 0; // a number to operate on so it is not optimized away
foreach (string stringWithCommaSeparatedNumbers in _stringsWithCommaSeparatedNumbers)
{
totalNumberOfIds += SplitToStackSpansWithEmptyCheck(stringWithCommaSeparatedNumbers).Length;
}
return totalNumberOfIds;
}
private static int[] SplitToStackSpansWithEmptyCheck(string path)
{
ReadOnlySpan<char> pathSpan = path.AsSpan();
MemoryExtensions.SpanSplitEnumerator<char> pathSegments = pathSpan.Split(Constants.CharArrays.Comma);
// Here we do NOT know how large the resulting list should at least be
// Default empty List<> internal array capacity on add is currently 4
// If the count of numbers are less than 4, we overallocate a little
// If the count of numbers are more than 4, the list will be resized, resulting in a copy from initial array to new double size array
// If the count of numbers are more than 8, another new array is allocated and copied to
List<int> nodeIds = [];
foreach (Range rangeOfPathSegment in pathSegments)
{
// this is only a span of the string, a string is not allocated on the heap
ReadOnlySpan<char> pathSegmentSpan = pathSpan[rangeOfPathSegment];
if (pathSegmentSpan.IsEmpty)
{
continue;
}
if (int.TryParse(pathSegmentSpan, NumberStyles.Integer, CultureInfo.InvariantCulture, out int pathSegment))
{
nodeIds.Add(pathSegment);
}
}
var result = new int[nodeIds.Count];
var resultIndex = 0;
for (int i = nodeIds.Count - 1; i >= 0; i--)
{
result[resultIndex++] = nodeIds[i];
}
return result;
}
// BenchmarkDotNet v0.14.0, Windows 11 (10.0.26100.2894)
// Intel Core i7-10750H CPU 2.60GHz, 1 CPU, 12 logical and 6 physical cores
// .NET Core SDK 3.1.426 [C:\Program Files\dotnet\sdk]
// [Host] : .NET 9.0.0 (9.0.24.52809), X64 RyuJIT AVX2
//
// Toolchain=InProcessEmitToolchain
//
// | Method | Mean | Error | StdDev | Gen0 | Allocated |
// |------------------------------------------------------ |---------:|---------:|---------:|-------:|----------:|
// | Linq | 46.39 us | 0.515 us | 0.430 us | 9.4604 | 58.31 KB |
// | SplitToHeapStrings | 30.28 us | 0.310 us | 0.275 us | 7.0801 | 43.55 KB |
// | SplitToStackSpansWithoutEmptyCheckReversingListAsSpan | 20.47 us | 0.290 us | 0.257 us | 2.7161 | 16.73 KB |
// | SplitToStackSpansWithoutEmptyCheck | 20.60 us | 0.315 us | 0.280 us | 2.7161 | 16.73 KB |
// | SplitToStackSpansWithEmptyCheck | 20.57 us | 0.308 us | 0.288 us | 2.7161 | 16.73 KB |
}