comments

pa_job_manager.py

@@ -1128,6 +1128,7 @@ def GetDateFromFile(file, stat):
             check2 = datetime( year, month, day )
         # give up and use file sys date
         except:
+            # use mtime - probably best to based date from when the content was created/changed 
             year, month, day, _, _, _, _, _, _ = datetime.fromtimestamp(stat.st_mtime).timetuple()
     c=date(year, month, day).isocalendar()
     woy=c[1]
@@ -1206,6 +1207,7 @@ def JobImportDir(job):
             fname=dir.PathOnFS()+'/'+basename
 
             stat = os.stat(fname)
+            # use ctime as even a metadata change (mv'd file on the fs, or a perms change) needs to be checked
             if stat.st_ctime > dir.last_import_date:
                 if DEBUG:
                     print("DEBUG: {} - {} is newer than {}".format( basename, stat.st_ctime, dir.last_import_date ) )
@@ -1349,6 +1351,7 @@ def GenHashAndThumb(job, e):
     if job.current_file_num % 100 == 0:
         session.commit()
     stat = os.stat( e.FullPathOnFS() )
+    # use mtime as only if the content is different do we need to redo the hash
     if stat.st_mtime < e.file_details.last_hash_date:
         if DEBUG:
             print(f"OPTIM: GenHashAndThumb {e.name} file is older than last hash, skip this")