perf(blame): run blame for entire file instead of per line

Previously current_line_blame would run a git-blame process per line
(via the `-L` flag) in an attempt to be more efficient. However after
some investigation it seems that running git-blame for the entire file
rarely exceeds 2x the time it takes to run for a single line, even for
large files.

This change alters current_line_blame to run git-blame for the entire
file after each buffer edit and caches that in memory. This makes the
first git-blame after an edit ~2x slower, but makes any cursor movements
after that instant.

A follow-up to this would be for current_line_blame to track buffer
updates to avoid the cache needing to be invalidated on every edit.

doc/gitsigns.txt

@@ -828,6 +828,7 @@ current_line_blame_formatter    *gitsigns-config-current_line_blame_formatter*
                          • `summary`: string
                          • `previous`: string
                          • `filename`: string
+                         • `boundary`: true?
 
                        Note that the keys map onto the output of:
                          `git blame --line-porcelain`

lua/gitsigns/actions.lua

@@ -813,7 +813,7 @@ end
 local function get_blame_hunk(repo, info)
   local a = {}
   -- If no previous so sha of blame added the file
-  if info.previous then
+  if info.previous_sha and info.previous_filename then
     a = repo:get_show_text(info.previous_sha .. ':' .. info.previous_filename)
   end
   local b = repo:get_show_text(info.sha .. ':' .. info.filename)
@@ -884,12 +884,14 @@ M.blame_line = async.void(function(opts)
   local buftext = util.buf_lines(bufnr)
   local fileformat = vim.bo[bufnr].fileformat
   local lnum = api.nvim_win_get_cursor(0)[1]
-  local result = bcache.git_obj:run_blame(buftext, lnum, opts.ignore_whitespace)
+  local results = bcache.git_obj:run_blame(buftext, lnum, opts.ignore_whitespace)
   pcall(function()
     loading:close()
   end)
 
-  assert(result)
+  assert(results and results[lnum])
+
+  local result = util.convert_blame_info(results[lnum])
 
   local is_committed = result.sha and tonumber('0x' .. result.sha) ~= 0
 

lua/gitsigns/attach.lua

@@ -337,6 +337,7 @@ local attach_throttled = throttle_by_id(function(cbuf, ctx, aucmd)
   end
 
   cache[cbuf] = CacheEntry.new({
+    bufnr = cbuf,
     base = ctx and ctx.base or config.base,
     file = file,
     commit = commit,

lua/gitsigns/cache.lua

@@ -7,6 +7,7 @@ local M = {
 -- Timer object watching the gitdir
 
 --- @class Gitsigns.CacheEntry
+--- @field bufnr              integer
 --- @field file               string
 --- @field base?              string
 --- @field compare_text?      string[]

lua/gitsigns/config.lua

@@ -658,6 +658,7 @@ M.schema = {
                          • `summary`: string
                          • `previous`: string
                          • `filename`: string
+                         • `boundary`: true?
 
                        Note that the keys map onto the output of:
                          `git blame --line-porcelain`

lua/gitsigns/current_line_blame.lua

@@ -20,52 +20,12 @@ local function reset(bufnr)
   vim.b[bufnr].gitsigns_blame_line_dict = nil
 end
 
--- TODO: expose as config
-local max_cache_size = 1000
-
---- @class Gitsigns.BlameCache
---- @field cache table<integer,Gitsigns.BlameInfo>
---- @field size integer
+--- @class (exact) Gitsigns.BlameCache
+--- @field cache Gitsigns.BlameInfo[]?
 --- @field tick integer
 
-local BlameCache = {}
-
 --- @type table<integer,Gitsigns.BlameCache>
-BlameCache.contents = {}
-
---- @param bufnr integer
---- @param lnum integer
---- @param x? Gitsigns.BlameInfo
-function BlameCache:add(bufnr, lnum, x)
-  if not x then
-    return
-  end
-  if not config._blame_cache then
-    return
-  end
-  local scache = self.contents[bufnr]
-  if scache.size <= max_cache_size then
-    scache.cache[lnum] = x
-    scache.size = scache.size + 1
-  end
-end
-
---- @param bufnr integer
---- @param lnum integer
---- @return Gitsigns.BlameInfo?
-function BlameCache:get(bufnr, lnum)
-  if not config._blame_cache then
-    return
-  end
-
-  -- init and invalidate
-  local tick = vim.b[bufnr].changedtick
-  if not self.contents[bufnr] or self.contents[bufnr].tick ~= tick then
-    self.contents[bufnr] = { tick = tick, cache = {}, size = 0 }
-  end
-
-  return self.contents[bufnr].cache[lnum]
-end
+local blame_cache = {}
 
 --- @param fmt string
 --- @param name string
@@ -93,24 +53,38 @@ end
 --- @param opts Gitsigns.CurrentLineBlameOpts
 --- @return Gitsigns.BlameInfo?
 local function run_blame(bufnr, lnum, opts)
-  local result = BlameCache:get(bufnr, lnum)
+  -- init and invalidate
+  local tick = vim.b[bufnr].changedtick
+  if not blame_cache[bufnr] or blame_cache[bufnr].tick ~= tick then
+    blame_cache[bufnr] = { tick = tick }
+  end
+
+  local result = blame_cache[bufnr].cache
+
   if result then
-    return result
+    return result[lnum]
   end
 
   local buftext = util.buf_lines(bufnr)
   local bcache = cache[bufnr]
-  result = bcache.git_obj:run_blame(buftext, lnum, opts.ignore_whitespace)
-  BlameCache:add(bufnr, lnum, result)
+  result = bcache.git_obj:run_blame(buftext, nil, opts.ignore_whitespace)
+
+  if not result then
+    return
+  end
 
-  return result
+  blame_cache[bufnr].cache = result
+
+  return result[lnum]
 end
 
 --- @param bufnr integer
 --- @param lnum integer
 --- @param blame_info Gitsigns.BlameInfo
 --- @param opts Gitsigns.CurrentLineBlameOpts
 local function handle_blame_info(bufnr, lnum, blame_info, opts)
+  blame_info = util.convert_blame_info(blame_info)
+
   vim.b[bufnr].gitsigns_blame_line_dict = blame_info
 
   local bcache = assert(cache[bufnr])

lua/gitsigns/git.lua

@@ -14,6 +14,7 @@ local uv = vim.loop
 local startswith = vim.startswith
 
 local dprint = require('gitsigns.debug.log').dprint
+local dprintf = require('gitsigns.debug.log').dprintf
 local eprint = require('gitsigns.debug.log').eprint
 local err = require('gitsigns.message').error
 
@@ -564,13 +565,7 @@ Obj.unstage_file = function(self)
   self:command({ 'reset', self.file })
 end
 
---- @class Gitsigns.BlameInfo
---- -- Info in header
---- @field sha string
---- @field abbrev_sha string
---- @field orig_lnum integer
---- @field final_lnum integer
---- Porcelain fields
+--- @class Gitsigns.CommitInfo
 --- @field author string
 --- @field author_mail string
 --- @field author_time integer
@@ -580,46 +575,57 @@ end
 --- @field committer_time integer
 --- @field committer_tz string
 --- @field summary string
---- @field previous string
---- @field previous_filename string
---- @field previous_sha string
---- @field filename string
----
---- Custom fields
+--- @field sha string
+--- @field abbrev_sha string
+--- @field boundary? true
+
+--- @class Gitsigns.BlameInfoPublic: Gitsigns.BlameInfo, Gitsigns.CommitInfo
 --- @field body? string[]
 --- @field hunk_no? integer
 --- @field num_hunks? integer
 --- @field hunk? string[]
 --- @field hunk_head? string
 
+--- @class Gitsigns.BlameInfo
+--- @field orig_lnum integer
+--- @field final_lnum integer
+--- @field commit Gitsigns.CommitInfo
+--- @field filename string
+--- @field previous_filename? string
+--- @field previous_sha? string
+
+local NOT_COMMITTED = {
+  author = 'Not Committed Yet',
+  ['author_mail'] = '<not.committed.yet>',
+  committer = 'Not Committed Yet',
+  ['committer_mail'] = '<not.committed.yet>',
+}
+
+---@param x any
+---@return integer
+local function asinteger(x)
+  return assert(tonumber(x))
+end
+
 --- @param lines string[]
---- @param lnum integer
---- @param ignore_whitespace boolean
---- @return Gitsigns.BlameInfo?
+--- @param lnum? integer
+--- @param ignore_whitespace? boolean
+--- @return Gitsigns.BlameInfo[]?
 function Obj:run_blame(lines, lnum, ignore_whitespace)
-  local not_committed = {
-    author = 'Not Committed Yet',
-    ['author_mail'] = '<not.committed.yet>',
-    committer = 'Not Committed Yet',
-    ['committer_mail'] = '<not.committed.yet>',
-  }
-
   if not self.object_name or self.repo.abbrev_head == '' then
     -- As we support attaching to untracked files we need to return something if
     -- the file isn't isn't tracked in git.
     -- If abbrev_head is empty, then assume the repo has no commits
-    return not_committed
+    return NOT_COMMITTED
   end
 
-  local args = {
-    'blame',
-    '--contents',
-    '-',
-    '-L',
-    lnum .. ',+1',
-    '--line-porcelain',
-    self.file,
-  }
+  local args = { 'blame', '--contents', '-', '--incremental' }
+
+  if lnum then
+    vim.list_extend(args, { '-L', lnum..',+1' })
+  end
+
+  args[#args+1] = self.file
 
   if ignore_whitespace then
     args[#args + 1] = '-w'
@@ -634,36 +640,89 @@ function Obj:run_blame(lines, lnum, ignore_whitespace)
   if #results == 0 then
     return
   end
-  local header = vim.split(table.remove(results, 1), ' ')
-
-  --- @diagnostic disable-next-line:missing-fields
-  local ret = {} --- @type Gitsigns.BlameInfo
-  ret.sha = header[1]
-  ret.orig_lnum = tonumber(header[2]) --[[@as integer]]
-  ret.final_lnum = tonumber(header[3]) --[[@as integer]]
-  ret.abbrev_sha = string.sub(ret.sha, 1, 8)
-  for _, l in ipairs(results) do
-    if not startswith(l, '\t') then
-      local cols = vim.split(l, ' ')
-      --- @type string
-      local key = table.remove(cols, 1):gsub('-', '_')
-      --- @diagnostic disable-next-line:no-unknown
-      ret[key] = table.concat(cols, ' ')
-      if key == 'previous' then
-        ret.previous_sha = cols[1]
-        ret.previous_filename = cols[2]
+
+  local ret = {} --- @type Gitsigns.BlameInfo[]
+  local commits = {} --- @type table<string,Gitsigns.CommitInfo>
+  local i = 1
+
+  while i <= #results do
+    --- @param pat? string
+    --- @return string
+    local function get(pat)
+      local l = assert(results[i])
+      i = i + 1
+      if pat then
+        return l:match(pat)
       end
+      return l
     end
-  end
 
-  -- New in git 2.41:
-  -- The output given by "git blame" that attributes a line to contents
-  -- taken from the file specified by the "--contents" option shows it
-  -- differently from a line attributed to the working tree file.
-  if ret.author_mail == '<external.file>' or ret.author_mail == 'External file (--contents)' then
-    ret = vim.tbl_extend('force', ret, not_committed)
+    local function peek(pat)
+      local l = results[i]
+      if l and pat then
+        return l:match(pat)
+      end
+      return l
+    end
+
+    local sha, orig_lnum_str, final_lnum_str, size_str = get('(%x+) (%d+) (%d+) (%d+)')
+    local orig_lnum = asinteger(orig_lnum_str)
+    local final_lnum = asinteger(final_lnum_str)
+    local size = asinteger(size_str)
+
+    if peek():match('^author ') then
+      --- @type table<string,string|true>
+      local commit = {
+        sha = sha,
+        abbrev_sha = sha:sub(1, 8),
+      }
+
+      -- filename terminates the entry
+      while peek() and not peek():match('^filename ') do
+        local l = get()
+        local key, value = l:match('^([^%s]+) (.*)')
+        if key then
+          key = key:gsub('%-', '_') --- @type string
+          commit[key] = value
+        else
+          commit[l] = true
+          if l ~= 'boundary' then
+            dprintf("Unknown tag: '%s'", l)
+          end
+        end
+      end
+
+      -- New in git 2.41:
+      -- The output given by "git blame" that attributes a line to contents
+      -- taken from the file specified by the "--contents" option shows it
+      -- differently from a line attributed to the working tree file.
+      if commit.author_mail == '<external.file>' or commit.author_mail == 'External file (--contents)' then
+        commit = vim.tbl_extend('force', commit, NOT_COMMITTED)
+      end
+      commits[sha] = commit
+    end
+
+    local previous_sha, previous_filename = peek():match('^previous (%x+) (.*)')
+    if previous_sha then
+      get()
+    end
+
+    local filename = assert(get():match('^filename (.*)'))
+
+    for j = 0, size - 1 do
+      ret[final_lnum + j] = {
+        final_lnum = final_lnum + j,
+        orig_lnum = orig_lnum + j,
+        commit = commits[sha],
+        filename = filename,
+        previous_filename = previous_filename,
+        previous_sha = previous_filename
+      }
+    end
   end
 
+  assert(vim.tbl_isarray(ret))
+
   return ret
 end
 

lua/gitsigns/util.lua

@@ -243,4 +243,13 @@ function M.bufexists(buf)
   return vim.fn.bufexists(buf) == 1
 end
 
+--- @param x Gitsigns.BlameInfo
+--- @return Gitsigns.BlameInfoPublic
+function M.convert_blame_info(x)
+  --- @type Gitsigns.BlameInfoPublic
+  local ret = vim.tbl_extend('error', x, x.commit)
+  ret.commit = nil
+  return ret
+end
+
 return M