feat(diff): use diff-match-patch for better diff handling

README.md

@@ -514,6 +514,14 @@ make test
 
 See [CONTRIBUTING.md](/CONTRIBUTING.md) for detailed guidelines.
 
+# Acknowledgments
+
+## diff-match-patch
+
+CopilotChat.nvim includes [diff-match-patch (Lua port)](https://github.com/google/diff-match-patch) for diffing and patching functionality.  
+Copyright 2018 The diff-match-patch Authors.  
+Licensed under the Apache License 2.0.
+
 # Contributors
 
 Thanks goes to these wonderful people ([emoji key](https://allcontributors.org/docs/en/emoji-key)):

lua/CopilotChat/config/mappings.lua

@@ -172,11 +172,7 @@ return {
 
       local path = block.header.filename
       local bufnr = prepare_diff_buffer(path, source)
-      local new_lines, applied = diff.apply_diff(block, bufnr)
-      if not applied then
-        new_lines = utils.split_lines(block.content)
-      end
-
+      local new_lines = diff.apply_diff(block, bufnr)
       vim.api.nvim_buf_set_lines(bufnr, 0, -1, false, new_lines)
       local first, last = diff.get_diff_region(block, bufnr)
       if first and last then
@@ -227,15 +223,11 @@ return {
 
       local path = block.header.filename
       local bufnr = prepare_diff_buffer(path, source)
-      local new_lines, applied = diff.apply_diff(block, bufnr)
-      if not applied then
-        new_lines = utils.split_lines(block.content)
-      end
-      local original_lines = vim.api.nvim_buf_get_lines(bufnr, 0, -1, false)
+      local new_lines = diff.apply_diff(block, bufnr)
 
       local opts = {
         filetype = vim.bo[bufnr].filetype,
-        text = applied and table.concat(new_lines, '\n') or table.concat(original_lines, '\n'),
+        text = table.concat(new_lines, '\n'),
       }
 
       opts.on_show = function()

lua/CopilotChat/config/prompts.lua

@@ -29,7 +29,7 @@ The user is currently in workspace directory {DIR} (project root). File paths ar
 Context is provided to you in several ways:
 - Resources: Contextual data shared via "# <uri>" headers and referenced via "##<uri>" links
 - Code blocks with file path labels and line numbers (e.g., ```lua path=/file.lua start_line=1 end_line=10```)
-  Note: Line numbers prefixed to each line are for reference only and should never be included when outputting code
+  Note: Each line in code block can be prefixed with <line_number>: for your reference only. NEVER include these line numbers in your responses.
 - Visual selections: Text selected in visual mode that can be shared as context
 - Diffs: Changes shown in unified diff format (+, -, etc.)
 - Conversation history
@@ -41,7 +41,7 @@ If you can infer the project type (languages, frameworks, libraries) from contex
 For implementing features, break down the request into concepts and provide a clear solution.
 Think creatively to provide complete solutions based on the information available.
 Never fabricate or hallucinate file contents you haven't actually seen in the provided context.
-When outputting code, never include line number prefixes - they are only for reference when analyzing the provided context.
+When outputting code or diffs, NEVER include line number prefixes - they are only for reference when analyzing the provided context.
 </instructions>
 ]],
   },

lua/CopilotChat/instructions/edit_file_unified.lua

@@ -2,67 +2,33 @@ return [[
 <editFileInstructions>
 Return edits similar to unified diffs that `diff -U0` would produce.
 
-- Always include the first 2 lines with the file paths (no timestamps).
-- Start each hunk of changes with a `@@ ... @@` line.
-- Do not include line numbers in the hunk header.
-- The user's patch tool needs CORRECT patches that apply cleanly against the current contents of the file.
-- Indentation matters in the diffs!
+Make sure you include the first 2 lines with the file paths.
+Don't include timestamps with the file paths.
+Do not use any file path prefixes, just use --- path/to/file and +++ path/to/file.
 
-Context lines:
-- For each hunk that contains changes, you MUST always include 2-3 context lines before the change.
-- ALWAYS prefix every context line with a single space character.
-- Context lines MUST ONLY appear BEFORE changes, NEVER after changes.
-- MISSING CONTEXT LINES WILL CAUSE PATCH FAILURES - they are mandatory, not optional.
-- MISSING SPACE PREFIXES WILL CAUSE PATCH FAILURES - they are mandatory, not optional.
+Start each hunk of changes with a `@@` line.
 
-Change lines:
-- Mark all lines to be removed or changed with `-`.
-- Mark all new or modified lines with `+`.
-- Only output hunks that specify changes with `+` or `-` lines.
+The user's patch tool needs CORRECT patches that apply cleanly against the current contents of the file!
+Code can start with line number prefixes for reference (e.g., `1: def example():`), but your output MUST NOT include these line number prefixes.
+Think carefully and make sure you include and mark all lines that need to be removed or changed as `-` lines.
+Make sure you mark all new or modified lines with `+`.
+Don't leave out any lines or the diff patch won't apply correctly.
 
-Other instructions:
-- Start a new hunk for each section of the file that needs changes.
-- When editing a function, method, loop, etc., replace the entire code block: delete the entire existing version with `-` lines, then add the new, updated version with `+` lines.
-- To move code within a file, use 2 hunks: one to delete it from its current location, one to insert it in the new location.
-- To make a new file, show a diff from `--- /dev/null` to `+++ path/to/new/file.ext`.
+Indentation matters in the diffs!
 
-Example:
+Start a new hunk for each section of the file that needs changes.
 
-```diff
---- mathweb/flask/app.py
-+++ mathweb/flask/app.py
-@@ ... @@
--class MathWeb:
-+import sympy
-+
-+class MathWeb:
-@@ ... @@
--def is_prime(x):
--    if x < 2:
--        return False
--    for i in range(2, int(math.sqrt(x)) + 1):
--        if x % i == 0:
--            return False
--    return True
-@@ ... @@
--@app.route('/prime/<int:n>')
--def nth_prime(n):
--    count = 0
--    num = 1
--    while count < n:
--        num += 1
--        if is_prime(num):
--            count += 1
--    return str(num)
-+@app.route('/prime/<int:n>')
-+def nth_prime(n):
-+    count = 0
-+    num = 1
-+    while count < n:
-+        num += 1
-+        if sympy.isprime(num):
-+            count += 1
-+    return str(num)
-```
+Only output hunks that specify changes with `+` or `-` lines.
+
+Output hunks in whatever order makes the most sense.
+Hunks don't need to be in any particular order.
+
+When editing a function, method, loop, etc use a hunk to replace the *entire* code block.
+Delete the entire existing version with `-` lines and then add a new, updated version with `+` lines.
+This will help you generate correct code and correct diffs.
+
+To move code within a file, use 2 hunks: 1 to delete it from its current location, 1 to insert it in the new location.
+
+To make a new file, show a diff from `--- /dev/null` to `+++ path/to/new/file.ext`.
 </editFileInstructions>
 ]]