Use different GitHub API call to circumvent 300 file limit on PRs (#68)

Author: NikkelM

CHANGELOG.md

@@ -1,11 +1,15 @@
 # Changelog
 
-## v1.4.1
+## v1.4.2
 
 <!--Releasenotes start-->
-- Fixed the default value of the `additional_lines` configuration option not being set to 1.
+- Fixed the app not being able to run on PRs with more than 300 changed files.
 <!--Releasenotes end-->
 
+## v1.4.1
+
+- Fixed the default value of the `additional_lines` configuration option not being set to 1.
+
 ## v1.4.0
 
 - Added a new `additional_lines` configuration option to control how many lines below action items are rendered in code snippets.

README.md

@@ -42,6 +42,7 @@ To configure options, add a `todo-pr-checker` key at the top-level of your `.git
 ```yaml
 todo-pr-checker:
   post_comment: 'items_found'
+  ignore_files: ['testFolder/', '*.js']
   (...)
 ```
 

app.rb

@@ -236,48 +236,50 @@ def get_app_options(full_repo_name, head_sha)
 
   # (4) Retrieves all changes in a pull request from the GitHub API and formats them to be usable by the app
   def get_pull_request_changes(full_repo_name, pull_number, ignore_files_regex)
-    diff = @installation_client.pull_request(full_repo_name, pull_number, accept: 'application/vnd.github.diff')
-
     ignore_files_regex.map! do |pattern|
       pattern.gsub!('.', '\.')
       pattern.gsub!('*', '.*')
       pattern.gsub!('/', '\/')
       Regexp.new(pattern)
     end
 
-    current_file = ''
-    line_number = 0
     changes = {}
     ignored_files = []
 
-    diff_enum = diff.each_line
-    line = diff_enum.next rescue nil
+    # Paginate through all files in the pull request
+    page = 1
+    per_page = 100
     loop do
-      break if line.nil?
-
-      if line.start_with?('+++')
-        while line&.start_with?('+++')
-          current_file = line[6..].strip
-          if ignore_files_regex.any? { |pattern| pattern.match?(current_file) }
-            ignored_files << current_file
-            loop do
-              line = diff_enum.next rescue nil
-              break if line.nil? || line.start_with?('+++')
-            end
-          else
-            changes[current_file] = []
-            break
+      files = @installation_client.pull_request_files(full_repo_name, pull_number, per_page: per_page, page: page)
+      break if files.empty?
+
+      files.each do |file|
+        filename = file.filename
+        if ignore_files_regex.any? { |pattern| pattern.match?(filename) }
+          ignored_files << filename
+          next
+        end
+
+        changes[filename] = []
+        # Parse the patch to get added lines and their line numbers
+        next unless file.patch
+        line_number = 0
+        file.patch.each_line do |line|
+          if line.start_with?('@@')
+            # Extract the new file's starting line number from the hunk header
+            # Example: @@ -1,6 +1,7 @@
+            m = line.match(/\+([0-9]+)/)
+            line_number = m ? m[1].to_i - 1 : line_number
+          elsif line.start_with?('+') && !line.start_with?('+++')
+            changes[filename] << { line: line_number, text: line[1..] }
+            line_number += 1
+          elsif !line.start_with?('-') && !line.chomp.eql?('\ No newline at end of file')
+            line_number += 1
           end
         end
-        break if line.nil?
-      elsif line.start_with?('+')
-        changes[current_file] << { line: line_number, text: line[1..] }
-      elsif line.start_with?('@@')
-        line_number = line.split()[2].split(',')[0].to_i - 1
       end
 
-      line_number += 1 unless line.start_with?('-') || line.chomp == '\ No newline at end of file'
-      line = diff_enum.next rescue nil
+      page += 1
     end
 
     [changes, ignored_files]

version.rb

@@ -1,3 +1,3 @@
 # frozen_string_literal: true
 
-VERSION = '1.4.1'
+VERSION = '1.4.2'