Refactor history management functions into their own file

autoload/ale/engine.vim

@@ -9,30 +9,6 @@
 " output: The array of lines for the output of the job.
 let s:job_info_map = {}
 
-function! ale#engine#AddToHistory(buffer, status, job_id, command) abort
-    if g:ale_max_buffer_history_size <= 0
-        " Don't save anything if the history isn't a positive number.
-        let g:ale_buffer_info[a:buffer].history = []
-
-        return
-    endif
-
-    let l:history = g:ale_buffer_info[a:buffer].history
-
-    " Remove the first item if we hit the max history size.
-    if len(l:history) >= g:ale_max_buffer_history_size
-        let l:history = l:history[1:]
-    endif
-
-    call add(l:history, {
-    \   'status': a:status,
-    \   'job_id': a:job_id,
-    \   'command': a:command,
-    \})
-
-    let g:ale_buffer_info[a:buffer].history = l:history
-endfunction
-
 function! s:GetJobID(job) abort
     if has('nvim')
         "In NeoVim, job values are just IDs.
@@ -471,7 +447,7 @@ function! s:RunJob(options) abort
         " Add the job to the list of jobs, so we can track them.
         call add(g:ale_buffer_info[l:buffer].job_list, l:job)
 
-        let l:status = 'ran'
+        let l:status = 'started'
         let l:job_id = s:GetJobID(l:job)
         " Store the ID for the job in the map to read back again.
         let s:job_info_map[l:job_id] = {
@@ -482,7 +458,7 @@ function! s:RunJob(options) abort
         \}
     endif
 
-    call ale#engine#AddToHistory(l:buffer, l:status, l:job_id, l:command)
+    call ale#history#Add(l:buffer, l:status, l:job_id, l:command)
 endfunction
 
 " Determine which commands to run for a link in a command chain, or