Use a single TextEncoder instance

[grantcox]

This is a minor performance improvement. Currently a new TextEncoder instance is created for every JS string that is encoded into a byte array. Creating new instances every time is unnecessary AFAICT, and a single TextEncoder can be used for all conversions.

I've added a new benchmark to demonstrate the impact of this (see the bottom table row) - it's fairly minor, but measurable in our workflow.

(FYI this screenshot was taken on a branch that also had [PR 273](https://github.com//pull/273) applied, but I believe the relative improvement here is generally applicable)

Checklist

• I grant to recipients of this Project distribution a perpetual,
  non-exclusive, royalty-free, irrevocable copyright license to reproduce, prepare
  derivative works of, publicly display, sublicense, and distribute this
  Contribution and such derivative works.
• I certify that I am legally entitled to grant this license, and that this
  Contribution contains no content requiring a license from any third party.

[gemini-code-assist]

Hello @grantcox, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

Summary of Changes

This pull request introduces a performance optimization by using a single TextEncoder instance instead of creating a new one for each string encoding. It also adds a new benchmark to measure the impact of this change. The changes affect the sqlite-api.js file by introducing a textEncoder variable and using it in the createUTF8 function. Additionally, the benchmark suite is expanded to include a test case with 100K SELECT statements with bound string parameters, which required changes to benchmarks.html, benchmarks.js, and demo-worker.js.

Highlights

• Performance Optimization: The primary goal is to improve performance by reusing a single TextEncoder instance, reducing the overhead of creating new instances for each string encoding.
• New Benchmark: A new benchmark (Test 17) is added to specifically measure the performance impact of the TextEncoder optimization when using bound string parameters.
• Parameter Binding: The demo-worker.js file is modified to handle parameterized queries, allowing the new benchmark to execute SELECT statements with bound string parameters.

Changelog

Click here to see the changelog

• demo/benchmarks/benchmarks.html
  • Added a new table row for 'Test 17: 100K SELECTs with bound string parameters' at line 107.
• demo/benchmarks/benchmarks.js
  • Added a function benchmark100kSelectsWithParams to generate an array of 100,000 SELECT statements with parameters, starting at line 5.
  • Modified benchmarksReady to include the new benchmark with bound parameters, starting at line 15.
  • Renamed sql parameter to query in the query function at line 110.
• demo/demo-worker.js
  • Modified the message handler to process both single SQL strings and arrays of queries (SQL + params), starting at line 135.
  • Added logic to bind parameters to the prepared statement using sqlite3.bind_collection at line 150.
• src/sqlite-api.js
  • Introduced a textEncoder constant at line 37 to hold a single instance of TextEncoder.
  • Replaced new TextEncoder() with textEncoder in the createUTF8 function at line 42.
  • Replaced new TextEncoder() with textEncoder when encoding the SQL string to UTF-8 at line 663.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point in your pull request via creating an issue comment (i.e. comment on the pull request page) using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in issue comments and review comments.
Help	/gemini help	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist is currently in preview and may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments to provide feedback.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

---

A coder's quest for speed,
TextEncoder's wasteful deed,
One instance to bind,
A faster find,
Efficiency's planted seed.

Footnotes

1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

[gemini-code-assist]

Code Review

This pull request introduces a performance improvement by using a single TextEncoder instance instead of creating a new one for each string encoding. The change is well-explained and includes a benchmark demonstrating the impact. The code changes look good overall.

Summary of Findings

• Missing comments: Consider adding comments to explain the purpose of the string arguments in benchmark100kSelectsWithParams and the order in which benchmarks are loaded.
• Missing comment: Consider adding a comment to explain the performance benefits of initializing TextEncoder once.

Merge Readiness

The pull request is well-structured and addresses the stated goal of improving performance. The changes are relatively small and the inclusion of a benchmark is commendable. I recommend merging after addressing the comments above. I am unable to directly approve this pull request, so please have others review and approve this code before merging.

[gemini-code-assist]

This is a good optimization! Initializing TextEncoder once and reusing it avoids unnecessary object creation. Consider adding a comment explaining why this is being done, referencing the performance benefits.

  // Initialize TextEncoder once for performance to avoid repeated object creation
  const textEncoder = new TextEncoder();

[rhashimoto]

Thanks for the PR! I'm fine with the actual code change, but I don't want to add the extra benchmark as that set of benchmarks is based on another pre-existing set.

[grantcox]

No problem, I've dropped that commit now.