v2: Remove Custom field and enhance Extensions for custom elements

[mmcdole]

Summary

This PR removes the Item.Custom field entirely and enhances the existing Extensions field to handle all custom and non-standard elements with full structural support including attributes and nested elements. Custom element support has been added to both RSS and Atom parsers.

Changes

• Remove Custom field from gofeed.Feed and gofeed.Item structs
• Remove Custom field from rss.Item struct
• Update RSS parser to add non-namespaced elements to Extensions under "_custom" namespace
• Update Atom parser to add non-namespaced elements to Extensions under "_custom" namespace
• Add support for custom elements at both channel/feed and item/entry level in both formats
• Add convenient helper methods for accessing extensions
• Fix RSS parser to properly handle RDF structural elements
• Add comprehensive test coverage for both RSS and Atom custom elements

Breaking Changes

• Item.Custom field removed completely
• Feed.Custom field removed completely (if it existed)

Migration Guide

// Old way
value := item.Custom["myField"]

// New way - simple migration
value := item.GetCustomValue("myField")

// New way - with full access to attributes
exts := item.GetExtension("_custom", "myField")
if len(exts) > 0 {
    value = exts[0].Value
    attrs = exts[0].Attrs
}

Benefits

• Custom elements now support attributes (fixes the limitation of map[string]string)
• Custom elements can have nested structure
• Unified approach - all non-standard elements go through Extensions in both RSS and Atom
• Better namespace safety with "_custom" prefix to avoid conflicts
• Consistent behavior across all feed formats
• Better type safety with the Extension struct
• Backward compatibility through helper methods

Test Coverage

Added comprehensive tests for:

• RSS custom elements with attributes
• RSS custom elements alongside namespaced extensions
• RSS custom elements with CDATA content
• RSS multiple custom elements with the same name
• RSS feed-level custom elements
• Atom feed-level custom elements
• Atom entry-level custom elements
• Atom multiple custom elements with same name
• Helper method functionality

Issues Fixed

• Fixes v2: Implement Structured Extension/Custom Element Handling #246 (main issue for this work)
• Fixes Enhance custom element handling #205 (custom elements now support nested structure and attributes)
• Fixes Parsing custom tags in Item #82 (custom tags like <weight> are now accessible via GetCustomValue)

Example Usage

// Parse a feed with custom elements (works for both RSS and Atom)
feed, _ := parser.ParseURL("https://ctftime.org/event/list/upcoming/rss/")

// Access custom elements in RSS
for _, item := range feed.Items {
    weight := item.GetCustomValue("weight")           // "54.67"
    format := item.GetCustomValue("format_text")      // "Jeopardy" 
    startDate := item.GetCustomValue("start_date")    // "20250530T120000"
}

// Access custom elements in Atom
atomFeed, _ := parser.Parse(atomReader)
for _, entry := range atomFeed.Entries {
    priority := entry.GetCustomValue("priority")      // "1"
    customId := entry.GetCustomValue("customId")      // "entry-123"
    
    // Get with attributes
    priorityExt := entry.GetExtension("_custom", "priority")
    if len(priorityExt) > 0 {
        level := priorityExt[0].Attrs["level"]        // "high"
    }
}

[mmcdole]

@infogulch let me know if this change seems fine to you as well.

[infogulch]

Looks ok to me. Does this save pagination information like what's described at https://stackoverflow.com/questions/1301392/pagination-in-feeds-like-atom-and-rss ?

[mmcdole]

Does this save pagination information

I think this is still fine, if I understand the StackOverflow link correctly.

Looking at our Atom parser, we already capture link elements with their rel attributes in the Links field. So pagination links like:

  <link rel="next" href="http://example.org/index.atom?page=2"/>
  <link rel="previous" href="http://example.org/index.atom?page=1"/>
  <link rel="first" href="http://example.org/index.atom"/>
  <link rel="last" href="http://example.org/index.atom?page=10"/>

These get parsed into the atom.Feed.Links array, where each Link struct has the Href and Rel fields you'd need to follow pagination.

So you should be able to access pagination info like:

atomFeed, _ := parser.Parse(reader)
 for _, link := range atomFeed.Links {
     if link.Rel == "next" {
         nextPageURL := link.Href
         // fetch next page
     }
 }

This PR shouldn't change that, as we are only touching the custom fields and these are handled by the standard Link parsing.