perf(rspack_core): replace HASH_PLACEHOLDER_REGEX with simple string parsing

[shulaoda]

Summary

Related to rolldown/rolldown#1496

I plan to convert all the regular expressions into string lookup, because these regular expressions are very simple and do not require regex.

rspack/crates/rspack_core/src/options/filename.rs

Lines 19 to 38 in a06f078

	pub static FILE_PLACEHOLDER: LazyLock<Regex> =
	LazyLock::new(|| Regex::new(r"\[file\]").expect("Should generate regex"));
	pub static BASE_PLACEHOLDER: LazyLock<Regex> =
	LazyLock::new(|| Regex::new(r"\[base\]").expect("Should generate regex"));
	pub static NAME_PLACEHOLDER: LazyLock<Regex> =
	LazyLock::new(|| Regex::new(r"\[name\]").expect("Should generate regex"));
	pub static PATH_PLACEHOLDER: LazyLock<Regex> =
	LazyLock::new(|| Regex::new(r"\[path\]").expect("Should generate regex"));
	pub static EXT_PLACEHOLDER: LazyLock<Regex> =
	LazyLock::new(|| Regex::new(r"\[ext\]").expect("Should generate regex"));
	pub static QUERY_PLACEHOLDER: LazyLock<Regex> =
	LazyLock::new(|| Regex::new(r"\[query\]").expect("Should generate regex"));
	pub static FRAGMENT_PLACEHOLDER: LazyLock<Regex> =
	LazyLock::new(|| Regex::new(r"\[fragment\]").expect("Should generate regex"));
	pub static ID_PLACEHOLDER: LazyLock<Regex> =
	LazyLock::new(|| Regex::new(r"\[id\]").expect("Should generate regex"));
	pub static RUNTIME_PLACEHOLDER: LazyLock<Regex> =
	LazyLock::new(|| Regex::new(r"\[runtime\]").expect("Should generate regex"));
	pub static URL_PLACEHOLDER: LazyLock<Regex> =
	LazyLock::new(|| Regex::new(r"\[url\]").expect("Should generate regex"));

Checklist

• Tests updated (or not required).
• Documentation updated (or not required).

[netlify]

✅ Deploy Preview for rspack canceled.

Built without sensitive environment variables

Name	Link
🔨 Latest commit	a06f078
🔍 Latest deploy log	https://app.netlify.com/sites/rspack/deploys/670640b86100ef0008c9cd22

[chenjiahan]

@shulaoda can you resolve the conflicts? ❤️

[h-a-n-a]

Wasn't the issue in the description had been resolved? I think this is necessary.

I'm not sure if the hash pattern appears only once. If there are cases where the pattern appears multiple times, such as [hash]-[hash:6].js, I need to optimize further.

let result = replace_all_hash_pattern("hello-[hash]-[hash:5].js", "[hash]", |n| {
    &"abcdefgh"[..n.unwrap_or(8)]
  });
 assert_eq!(result, Some("hello-abcdefgh-abcde.js".to_string()));

[shulaoda]

Wasn't the issue in the description had been resolved? I think this is necessary.

Sorry. This is a PR from a long time ago, and I fixed the issues in the description later. I have revised the description.