Cascadingly capture css @import urls and extracting urls from separate css item

[yzqzss]

• Supported extracting URLs from individual embedded CSS items. ("embedded" means the CSS item is must derived from a HTML item.)
• Supported following CSS @import jumps, capturing @import URLs cascadingly.
• To prevent infinite @import jumps, added the --max-css-jump option to limit it.
• Since the tdewolff/parse parser currently does not support nested CSS syntax, the previous regex extractor was added back as a fallback when tdewolff/parse fails. (Also improved the previous CSS regex; it should now produce less non-existent links.)

---

close: #323

[Copilot]

Pull Request Overview

This PR adds support for cascading CSS @import handling by detecting embedded CSS items, extracting their URLs, and controlling import depth.

• Introduced detection of embedded CSS items and jump-count logic
• Hooked into HTML and CSS extraction to enqueue @import URLs as children
• Added MaxCSSJump configuration and CLI flag to limit import recursion

Reviewed Changes

Copilot reviewed 6 out of 6 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
internal/pkg/postprocessor/item.go	Skip depth limit for embedded CSS items when crawling assets
internal/pkg/postprocessor/extractor/html.go	Call AddAtImportLinksToItemChild after parsing CSS in HTML <style> blocks
internal/pkg/postprocessor/extractor/css_embedded.go	New logic for detecting embedded CSS and handling @import jump limits
internal/pkg/postprocessor/assets.go	Added branch for extracting assets from CSS items and capturing @import links
internal/pkg/config/config.go	Added MaxCSSJump field to configuration
cmd/get.go	Exposed --max-css-jump CLI flag

Comments suppressed due to low confidence (3)

internal/pkg/postprocessor/extractor/css_embedded.go:1

• There are no unit tests covering the new embedded CSS extractor logic. Please add tests for IsEmbeddedCSS, GetEmbeddedCSSJump, and AddAtImportLinksToItemChild to validate expected behavior.

package extractor

internal/pkg/postprocessor/item.go:74

• The word "embeded" is misspelled; please correct it to "embedded".

// 1. the item [is not an embeded css item] and [is a child has a depth (without redirections) bigger than 2].

internal/pkg/postprocessor/extractor/css_embedded.go:26

• The variable cssLogger is not defined or imported in this file, which will cause a compile error. Please import the appropriate logging package or reference the correct logger instance.

cssLogger.Warn("item is not an embedded CSS, returning 0 @import jump", "func", "GetEmbeddedCSSJump", "item_id", item.GetShortID())

[yzqzss]

The biggest issue with this CSS parser at the moment is that it doesn't support nested CSS yet, which is widely used today.

tdewolff/parse#92

Writing a fix for upstream might be time-consuming, so the mitigation is to add "if parsing fails, fall back to the regex based extractor we used before" logic. I'll to add the logic hours later...

[NGTmeaty]

As is, this looks really good! I'll keep my eyes out for the additional patch for nested CSS!

[yzqzss]

The parseErr include newlines, like this:

unexpected token '[' in declaration on line 5 and column 5
    5:     [aria-hidden="true"] {
           ^

So I put the err parameter at the end to make the terminal log look nicer.

[NGTmeaty]

This looks great! And thanks a ton for adding the regex fallback.