diff options
| author | erdgeist <erdgeist@erdgeist.org> | 2026-07-10 00:35:24 +0200 |
|---|---|---|
| committer | erdgeist <erdgeist@erdgeist.org> | 2026-07-10 00:35:24 +0200 |
| commit | a25d90335ad3738b6831288190132c2f7498465c (patch) | |
| tree | 5aa7083d9f8c8701644fe60bc7d28542775694b8 /lib/html_word_diff.rb | |
| parent | da0b79a7c1f3116e2b450ecbe50efdf31cf23414 (diff) | |
Retire vendored cacycle_diff.js for a diff-lcs-based diff view
Revisions#diff now computes an inline or side-by-side word diff
server-side (Page#diff_against, lib/html_word_diff.rb) instead of
shipping raw/escaped content to the browser for a 2008-era vendored
JS differ to mangle. View mode is picked via a `view` param, settable
either on the diff page itself or directly from the revisions#index
sticky bar next to "Diff revisions".
Also fixes a pre-existing bug in RevisionsController#diff's single-
revision branch, which set params[:start]/params[:end] instead of
params[:start_revision]/params[:end_revision].
Diffstat (limited to 'lib/html_word_diff.rb')
| -rw-r--r-- | lib/html_word_diff.rb | 52 |
1 files changed, 52 insertions, 0 deletions
diff --git a/lib/html_word_diff.rb b/lib/html_word_diff.rb new file mode 100644 index 0000000..1f28cf5 --- /dev/null +++ b/lib/html_word_diff.rb | |||
| @@ -0,0 +1,52 @@ | |||
| 1 | require 'diff/lcs' | ||
| 2 | |||
| 3 | module HtmlWordDiff | ||
| 4 | TOKEN_REGEXP = /<[^>]*>|[[:space:]]+|[[:alnum:]]+|[^[:space:][:alnum:]<>]/ | ||
| 5 | |||
| 6 | def self.inline(old_html, new_html) | ||
| 7 | html = +'' | ||
| 8 | each_change(old_html, new_html) do |action, old_token, new_token| | ||
| 9 | case action | ||
| 10 | when '=' | ||
| 11 | html << new_token | ||
| 12 | when '-' | ||
| 13 | html << "<del>#{old_token}</del>" | ||
| 14 | when '+' | ||
| 15 | html << "<ins>#{new_token}</ins>" | ||
| 16 | when '!' | ||
| 17 | html << "<del>#{old_token}</del><ins>#{new_token}</ins>" | ||
| 18 | end | ||
| 19 | end | ||
| 20 | html | ||
| 21 | end | ||
| 22 | |||
| 23 | def self.side_by_side(old_html, new_html) | ||
| 24 | old_out = +'' | ||
| 25 | new_out = +'' | ||
| 26 | each_change(old_html, new_html) do |action, old_token, new_token| | ||
| 27 | case action | ||
| 28 | when '=' | ||
| 29 | old_out << old_token | ||
| 30 | new_out << new_token | ||
| 31 | when '-' | ||
| 32 | old_out << "<del>#{old_token}</del>" | ||
| 33 | when '+' | ||
| 34 | new_out << "<ins>#{new_token}</ins>" | ||
| 35 | when '!' | ||
| 36 | old_out << "<del>#{old_token}</del>" | ||
| 37 | new_out << "<ins>#{new_token}</ins>" | ||
| 38 | end | ||
| 39 | end | ||
| 40 | [old_out, new_out] | ||
| 41 | end | ||
| 42 | |||
| 43 | def self.each_change(old_html, new_html) | ||
| 44 | Diff::LCS.sdiff(tokenize(old_html), tokenize(new_html)).each do |change| | ||
| 45 | yield change.action, change.old_element, change.new_element | ||
| 46 | end | ||
| 47 | end | ||
| 48 | |||
| 49 | def self.tokenize(html) | ||
| 50 | html.to_s.scan(TOKEN_REGEXP) | ||
| 51 | end | ||
| 52 | end | ||
